As the Internet has become a ubiquitous tool in modern science, it is increasingly important to evaluate the currency of free science information on the web. However, there are few empirical studies which have specifically focused on this issue. In this paper, we used the search engines Google, Yahoo and Altavista to generate a list of web pages about 32 terms. Sample pages were examined according to the criteria which were developed in this study. Results revealed that the mean of currency of free science information on the web was 2.6482 (n=2814), only 982 (34.90%) of samples got higher mean scores than the average. Sample pages with different domain names or subjects had difference with significance (P < 0.05). In conclusion, the currency of free science information on the web is unsatisfactory. The developed criteria set here could be a useful instrument for researchers and the public to assess information currency on the web by themselves.