Objective To investigate the association of microcystin (MC) in drinking water with the incidence of colorectal cancer. Methods The study was designed as a retrospective cohort. Eight townships or towns were randomly selected as the study sites in Haining City of Zhejiang Province, China. 408 cases of colon and rectum carcinomas diagnosed from 1977 to 1996 in the study sites were included, and a survey on types of drinking water of these patients was conducted. Samples of different water sources (well, tap, river and pond) were collected separately and microcystin concentrations were determined by indirect competitive ELISA method. Results The incidence rate of colorectal cancer was significantly higher in population who drank river and pond water than those who drank well and tap water. Compared to well water, the relative risk (RR) for colorectal cancer was 1.88 (tap), 7.94 (river) and 7.70 (pond) respectively. The positive rate (>50 pg/mL) of microcystin in samples of well, tap, river and pond water was 0, 0, 36.23% and 17.14% respectively. The concentration of microcystin in river and pond water was significantly higher than that in well and tap water (P<0.01). Spearman rank correlation analysis showed that in the study sites, the microcystin concentration of river and pond water was positively associated with the incidence of colorectal cancer (rs= 0.881, P<0.01). Conclusions The types of drinking water are positively associated with the incidence of colorectal cancer in the study sites, and this may be related to microcystin contamination of drinking water. Further biological study is needed to support the possible causative role of mycrocystin in carcinogenesis of colon and rectum. Objective To investigate the association of microcystin (MC) in drinking water with the incidence of colorectal cancer. Methods The study was designed as a retrospective cohort. Eight townships or towns were randomly selected as the study sites in Haining City of Zhejiang Province, China. 408 cases of colon and rectum carcinomas diagnosed from 1977 to 1996 in the study sites were included, and a survey on types of drinking water of these patients was conducted. Samples of different water sources (well, tap, river and pond) Results The incidence rate of colorectal cancer was significantly higher in population who drank river and pond water than those who drank well and tap water. Compared to well water, the relative risk (RR) for Colorectal cancer was 1.88 (tap), 7.94 (river) and 7.70 (pond) respectively. The positive rate (> 50 pg / mL) of microcystin in samples of well, tap , river and pond water was 0, 0, 36.23% and 17.14% respectively. The concentration of microcystin in river and pond water was significantly higher than that in well and tap water (P <0.01). Spearman rank correlation analysis showed that in the study sites, the microcystin concentration of river and pond water was positively associated with the incidence of colorectal cancer (rs = 0.881, P <0.01). Conclusions The types of drinking water are positively associated with the incidence of colorectal cancer in the study sites, and this may be related to microcystin contamination of drinking water. Further biological study is needed to support the possible causative role of mycrocystin in carcinogenesis of colon and rectum.