Abstract: Abstract: Productivity is one basic property of farm land and the spatial pattern can be used as the baseline information in making and implementing appropriate agricultural policies. As an important winter wheat production area of China, food production in the Huang-Huai-Hai region has been receiving considerable attention for a long time. The Normalized Difference Vegetation Index (NDVI) derived from remote sensing techniques is a widely utilized vegetation index in assessing agricultural land distribution, productivity and crop growth conditions. The NDVI time series can be disaggregated into a set of quantitative metrics reflecting crop distribution and growth phenology. However, few studies take both mean crop yield and its inter-annual variability into account in evaluating the productivity of wheat land. In the present paper, we proposed a method to evaluate wheat productivity by using annual NDVI indices derived from the time series of10 years MODIS NDVI data over the Huang-Huai-Hai region. The calculated productivity can be used to monitor farmland quality. The proposed method composed of four steps: Firstly, we analyzed the reconstructed NDVI time series to generate a set of phenology indices; Secondly we extracted the distribution of winter wheat in the study area using the decision tree classifier; Thirdly we ranked each growing season's mean NDVI(NDVImean) of winter wheat cultivated area from reviving to maturity stage into five levels and the Coefficient of Variation(C.V) of ten years NDVImean into four levels; Lastly, we evaluated the winter wheat land productivity by applying a criterion established by both mean rank and C.V level. The extracted wheat planted areas were consistent with the wheat sown area obtained from the statistic database at the county and provincial scale. The statistically significant correlation between the NDVImean and yields as calculated from the statistic database at the county levels suggested that the mean rank for ten years' NDVImean could indicate the yield level, whereas the C.V of the rank could measure the variation of yield within ten years. The relationship between the mean rank and the C.V for each 20 km×20 km grid showed that in most parts of the wheat land, the high ranked mean yield area had low temporal variations and the low rank's variations were high. It means the high variation of yield is related to a low productivity as well as a low yield level. Consequently, combining mean yield level and its inter-annual variability can generate more objective information for cropland productivity. The result showed that the percentage of high, middle and low level wheat land was 18.35%, 40.04% and 41.61% respectively, which suggested that there is still great potential to improve the winter wheat productivity in the study area. The spatial productivity pattern was driven by variations in hydrothermal condition, soil properties and water resources. The high productivity wheat lands were mainly located in the Huang-Huai Plain where hydrothermal condition is suitable for winter wheat. The low productivity lands were in Jin-Lu-Yu low Plain where drought is more likely to happen in spring, which is the key stage for winter wheat growth. Areas where wheat is low productive caused by infertile soil and water shortage, especially in Saline-alkali land, should be investigated in the future studies to help stabilize and increase grain yield in Huang-Huai-Hai region.