We construct a rice paddy biogeochemical cycle model to investigate processes governing rice mercury sources and to understand factors influencing spatiotemporal variability in Chinese rice mercury concentrations. The rice paddy model takes atmospheric mercury deposition, simulated from a global atmospheric chemistry transport model (GEOS‐Chem), and soil and irrigable surface water mercury concentrations obtained from literature and calculates rice inorganic (IHg) and methylmercury (MeHg) concentrations. We use ranges of GEOS‐Chem‐simulated future atmospheric mercury deposition—no policy and strict policy to regulate mercury emissions from Chinese coal‐fired power plants under the Minamata Convention on Mercury—to simulate future rice IHg and MeHg concentrations. Sensitivity analyses suggest that rice IHg and MeHg concentrations are more sensitive to the process of soil desorption than infiltration of recently introduced mercury (atmospheric and irrigation source). The rate of internal methylation via microbial activity has the largest modeled influence on rice MeHg concentration. We find that soil mercury, rather than atmospheric deposition, explains observed spatial variability in rice IHg and MeHg concentrations and captures locations of rice mercury hot spots (>20 ng/g; China National Standard Limit). Under our future scenarios, the Chinese median rice IHg and MeHg concentration increases by 13% and decrease by 18% under no policy and strict policy, respectively. Regions with the largest percentage decline in rice IHg and MeHg concentrations under strict policy are in central China, which have high rice mercury concentrations, rice production, and consumption. Our study suggests that addressing Chinese rice mercury contamination requires attention to contaminated soil and regulation of anthropogenic mercury emissions.