BackgroundAmbient fine particulate matter (PM2·5) pollution is becoming increasingly serious in China, and is a major risk factor for various cancers. Meanwhile, the regional transportation of air pollution is predicted to play an important role in terms of the effect of air pollution, against the background of massive internal migration. However, little study has focused on forecasting using long-term data and considering regional transportation. We aimed to predict spatial distribution of cancers related to PM2·5, namely breast cancer, pancreatic cancer, and all-cause cancer, and their mortalities and morbidities, and draw a series of forecasting maps of these cancers. MethodsWe included morbidity and mortality data from 1194 counties of breast cancer, pancreatic cancer, and all-cause cancer from 2006 to 2014. We also included the annual concentration of global surface PM2·5 concentration derived from satellite 0·01° × 0·01° spatial resolution. We used a spatial autocorrelation method to estimate the spatial relationship. A 1000-loops simulation was done to choose the optimal forecasting model between five alternative models: ridge regression, partial least square regression, regression tree, model tree, and the combined forecasting model. A kriging interpolation method was used to draw the distribution maps. FindingsThe trend showed a gradual increase in the mortality and morbidity of breast cancer, pancreatic cancer, and all-cause cancer. We found a significant spatial autocorrelation between cancer incidence and PM2·5. Our results from forecasting showed a constant growth in mortality and morbidity of all cancers, and the kriging method suggested a similar spatial pattern. High morbidity and mortality areas were mainly in central-east and south-east China. InterpretationWe found a similar distribution pattern between PM2·5 concentration and mortality and morbidity associated with PM2·5-related cancers. Our results serve as a valuable reference for the development of effective policies to reduce air pollution emissions, with the efforts from governments in high-risk areas. FundingNo funding.