The aim was to study the variation trends of all-cause and cancer mortality during 1984–2013 in Macheng City, China. The mortality data were collected from Macheng City disease surveillance points system and Hubei Center for Disease Control and Prevention. The model life table system was used to adjust mortality rates due to an under-reporting problem. An age-period-cohort model and intrinsic estimator algorithm were used to estimate the age effect, period effect, and cohort effect of all-cause mortality and cancer mortality for males and females. Age effect of all-cause mortality for both sexes increased with age, while the age effect of cancer mortality for both sexes reached a peak at the age group of 55–59 years old and then decreased. The relative risks (RRs) of all-cause mortality for males and females declined with the period and decreased by 51.13% and 63.27% during the whole study period, respectively. Furthermore, the period effect of cancer mortality in both sexes decreased at first and then increased. The cohort effect of all-cause and cancer mortality for both sexes born after 1904 presented the pattern of “rise first and then fall,” and decreased by 82.18% and 90.77% from cohort 1904–1908 to 1989–1993, respectively; especially, the risk of all-cause and cancer mortality for both sexes born before 1949 was much higher than that for those born after 1949.