Abstract: Plant type and irrigation scheme are key influencing factors of real-time yield estimation and monitoring of winter wheat in precision farming. In this paper, MODIS remote sensing data were used in combination with GPS and ground-truth non-remote sensing data to determine the dynamics of normalized difference vegetation index (NDVI) of winter wheat cultivars with different plant types under irrigation and non-irrigation conditions. The relationship between NDVI and yield of different winter wheat cultivars in different growth stages was then analyzed. Results showed the trends in NDVI with developmental stages of different wheat cultivars were same, following a low-high-low curve. There were obvious differences in NDVI from jointing to booting stages for different cultivars, and NDVI for cultivars with horizontal plant types was higher than that for cultivars with erect plant types. It implied that the jointing-to-booting stage was the best period for identifying plant types of winter wheat cultivars. Even for the same cultivar, mean NDVI was obviously different at each growth stage for irrigated and non-irrigated lands. NDVI for irrigated winter wheat was higher than that for non-irrigated winter wheat, with a notable difference especially at the early heading stage. At early heading stage, NDVI was strongly correlated with yield in irrigated and non-irrigated lands. However, regression equation based on NDVI both in the early heading and filling stages gave better prediction for wheat yield than that based on NDVI only in the early heading stage. This was especially the case for non-irrigated wheat fields.