This study described methods to predict human health risk associated with exposure to environmental carciongens using animal bioassay data. Also, biological assumption for various dose-response models were reviewed. To illustrate the process of risk estimate using relevant dose-response models such as Log-normal, Mantel-Bryan, Weibull and Multistage model, we used four animal carcinogenesis bioassy data of chloroform and chloroform concentrations of tap water measured in large cities of korea from 1987 to 1995. As a result, in the case of using average concentration in exposure data and 95 % upper boud unit risk of Multistage model, excess cancer risk(RISK I) was about 1.9 x 10-6, in the case of using probability distribution of cumulative exposure data and unit risks, those risks(RISK II) which were simulated by Monte-Carlo analysis were about 2.4 x 10(-6) and 7.9 x 10(-5) at 50 and 95 percentile, respectively. Therefore risk estimated by Monte-Carlo analysis using probability distribution of input variables may be more conservative.