Abstract: Food security is one of the main problems facing China and researches on establishing farmland productivity are of vital significance for food productivity and safety. Traditional calculations of farmland productivity have been based on farmland grading to establish functional relations among the natural quality of farmland, grade index, theoretical yield and achievable yield using sample surveys. The deficiencies that associated with this calculation method have included tedious data collection and strong subjectivity. In the attempt to avoid these deficiencies, this paper used vegetation net primary productivity (NPP) method to determine farmland productivity in Wuhan, Hubei Province. The paper used NDVI and EVI data, monthly average precipitation, monthly average temperature, monthly total solar radiation and other meteorological data to run the CASA model. The model then was used to calcu-late NPP of farmlands in the region and establish functional relations among theoretical and achievable yields, NPP of farmlands using sample survey data. The unit productivity (UP) was obtained by substituting NPP data into the established relation functions. Then productivities of different towns were aggregations of UP multipling its corresponding area, while the yields of towns were quotients of productivity divided by total farmland area of the towns. Theoretical and achievable productivities and yields in 75 townships in the study area were calculated using the NPP method. While townships with larger theoretical and achievable productivities were mainly in the northern and southeast regions, townships with larger theoretical and achievable yields were in the western and northeast regions of the study area. Also both theoretical and achievable productivities and yields were lower in peripheral townships in close proximity with main cities. By contrast, comparison of the NPP and traditional method showed that the calculated results by the two methods were highly consistent in terms of theoretical and achievable productivities, but not entirely consistent in terms of theoretical and achievable yields. The results therefore showed that it was feasible to apply NPP method on estimating the productivity of farmlands. The NPP method was advantageous over the traditional method because it not only avoided deficiencies inherent in the traditional method, but was also a quick way of calculating farmland productivity without need for farmland grading and rating.