CO2 fluxes and concentrations are not well understood in Northeast China, where dominant land surface types are mixed forest and cropland. Here, we analyzed the CO2 fluxes and concentrations using observations and the Weather Research and Forecasting model coupled with the Vegetation Photosynthesis and Respiration Model (WRF‐VPRM). We also used WRF‐VPRM outputs to examine CO2 transport/dispersion and budgets. Finally, we investigated the uncertainties of simulating CO2 fluxes related to four VPRM parameters (including maximum light use efficiency, photosynthetically active radiation half‐saturation value, and two respiration parameters) using off‐line ensemble simulations. The results indicated that mixed forests acted as a larger CO2 source and sink than rice paddies in 2016 due to a longer growth period and stronger ecosystem respiration, although measured minimum daily mean net ecosystem exchange (NEE) was smaller at rice paddy (−10 μmol·m‐2·s‐1) than at mixed forest (−6.5 μmol·m‐2·s‐1) during the growing season (May–September). The monthly fluctuation of column‐averaged CO2 concentrations (XCO2) exceeded 10 ppm in Northeast China during 2016. The large summertime biogenic sinks offset about 70% of anthropogenic contribution of XCO2 in this region. WRF‐VPRM modeling successfully captured seasonal and episodic variations of NEE and CO2 concentrations; however, NEE in mixed forest was overestimated during daytime, mainly due to the uncertainties of VPRM parameters, especially maximum light use efficiency. These findings suggest that the WRF‐VPRM modeling framework will provide greater understanding of the natural and anthropogenic contributions to the carbon cycle in China, especially after calibration of parameters that control biogenic fluxes.