The answer depends on what you call a "science." For example, if we use a broad definition supplied by the OED ("a particular branch of knowledge or study"), then law certainly qualifies.
If we adopt a more narrow definition of "science," (for instance, "results based on the scientific model") the answer gets murkier. Does law use the scientific method?
John Vielleux[1] argues that while the practice of law may not be perfectly scientific, there are similarities between the study of law and, say, biology. Legal decisions may not be universally applicable, but there is coherence in a specific body of jurisprudence. And although there are valid complaints about the political content of legal decisions, other sciences lack perfect objectivity as well.
Thomas Ulen[2] notes that law has historically emphasized "doctrinal scholarship:" the training of attorneys. However, legal scholars today are often doing interdisciplinary work, studying law using scientific methods from other social sciences (especially economics and sociology).
Further, Ulen points to a "growing disjunction" in law schools between doctrinal scholarship and scientific study of law. If this trend continues, we may see a separation between legal research and attorney training -- just as those who study business pursue PhDs in economics, but would-be CEOs earn MBAs.
[1]: John Villeux, "The Scientific Model in Law," 78 Georgetown Law Journal 1967 (1987)
[2]: Thomas S. Ulen, ""The Unexpected Guest: Law and Economics, Law and Other Cognate Disciplines, and the Future of Legal Scholarship." 79 Chicago-Kent Law Review 403 (2004)