Surely, A Theory of Justice is among the most important and influential texts in contemporary philosophy. And it is, of course, the central text in contemporary political philosophy. Want just a few reasons to think this is an important text that you ought to read? Here you go: Rawls develops and defends a new theory of justice, he provides a new way to extend some of the basic ideas in the social contract tradition, his text was crucial in resurrecting Kantian moral theory, his work has helped to bring constructivist meta-ethical positions back into prominence, the book develops some new and influential criticisms of utilitarianism, and it includes an explication of the method of reflective equilibrium and demonstrates how it can be applied in moral theory, etc.
This is a long, intricate, and densely argued book, and there's no hope of summarizing even its main claims in this review. Consequently, I'll simply aim to give a very sketchy account of the structure of his main argument here.
Rawls's theory is a theory of justice as it applies to the basic institutions of a single society. He calls his theory "justice as fairness." It is not that he thinks justice is simply fairness, or that a just society is a fair one. Rather, people choose principles of justice in a position that is supposed to be fair; their choices in this fair position determine the correct principles of justice. The principles of justice determine the nature of a just society; they apply to the basic structure of society--to its fundamental institutions. They will be understood by people who accept them as principles telling them how their society should be structured with respect to how it provides people with their basic rights and liberties, how it determines people's opportunities in life, and how it structures the institutions in which people acquire wealth and income.
The fair position for choosing these principles is what Rawls calls "the original position." His argument has the following structure: he describes the original position, and then he argues that parties in the original position would choose a particular set of principles of justice. The principles chosen constitute the correct theory of justice.
The first part of the argument is a detailed account of the original position. Parties in the original position are placed behind a veil of ignorance, where they are stripped of certain types of knowledge. In particular, they lose all the knowledge of the contingent facts concerning their own standing in life and the details of life in their society. Furthermore, they lose knowledge of their particular talents, desires, psychological traits, skills, etc. Why prefer this as a position in which principles of justice are to be chosen? The main idea is that it allows us to see the people as coming to fair terms for social cooperation, for this is supposed to be a fair situation for selecting the principles. Parties behind the veil are unable to rig the principles of justice to benefit themselves rather than others; they aren't allowed to use their position or talents to strongarm people into selecting principles that aren't to those people's benefit; and they aren't allowed to craft the principles to suit their actual needs, aims, desires, etc. However, parties in the original position do possess the sort of general knowledge about human psychology, human societies, and the natural world that would be required to choose between principles of justice.
Now, importantly, placing individuals in the original position depends on a particular moral view; this is supposed to reflect our considered judgments about justice and fairness. It is a way of drawing out what we actually think about these things. This is not a historical argument: the original position isn't supposed to be a description of some situation people were once in. Nor is this an argument grounded in some account of human nature and psychology: the parties in the original position aren't supposed to reflect something of importance about human psychology. (One should see section 40 for an account of this as a Kantian conception of justice, though. Here Rawls may be resting his theory on an account of us as beings of a certain sort. But, again, this is a philosophical and moral account of persons; this isn't the sort of thing you're going to find out about by doing ordinary sociology, anthropology, or psychology.)
In the next part of his argument Rawls claims that parties in the original position would agree upon the following principles of justice. The first principle is that individuals are to possess greatest amount of basic rights and liberties compatible with similar rights and liberties for others. The relevant rights and liberties are the right to vote and to hold public office, freedom of thought, freedom of speech and assembly, the right to own property and to avoid unreasonable search and seizure, etc. The second principle is that there is to be fair equality of opportunity with respect to positions of authority and responsibility, and that inequalities in wealth and income are be for the benefit of all, and particularly for the benefit of the worst-off group. The first principle is to be satisfied before the second one, so rights and liberties cannot be sacrificed in the interest of securing more wealth or income for any or all people. And one should notice that these principles do not clearly imply anything about how the institutions in which people acquire wealth and income are to be ordered or regulated. This will depend on which set of institutions would actually meet the requirements set by the second principles, and this will depend on empirical facts about how the world works. Moreover, it should be pointed out that many ways of ordering and regulating these institutions will be ruled out by the first principle, irrespective of how well off they would make the worst-off group.
This, clearly, should be read by anyone interested in contemporary analytic philosophy, and it is an absolutely crucial text for people studying ethics or political philosophy.