A study was made on cancer mortality in the high-background radiation areas of Yangjiang, China. Based on hamlet-specific environmental doses and sex- and age-specific occupancy factors, cumulative doses were calculated for each subject. In this article, we describe how the indirect estimation was made on individual dose and the methodology used to estimate radiation risk. Then, assuming a linear dose response relationship and using cancer mortality data for the period 1979-1995, we estimate the excess relative risk per Sievert for solid cancer to be -0.11 (95% CI, -0.67, 0.69). Also, we estimate the excess relative risks of four leading cancers in the study areas, i.e., cancers of the liver, nasopharynx, lung and stomach. In addition, we evaluate the effects of possible bias on our risk estimation.