The crop net primary productivity is inversion using HJ-1 CCD and IRS remote sensing images in Zhuozhou and Gaobeidian, Baoding City, Hebei Province. This inversion is done based on surface energy balance algorithm for land model. There are two important components in this inversion process: efficiency for solar energy utilization and absorbed photosynthetic active radiation (APAR). Firstly, HJ-1 CCD and IRS remote sensing images, DEM, meteorological data are used to estimate net radiation flux, soil heat flux, sensible heat flux to the air. These three fluxes are used to compute instantaneous evaporative fraction, which is used to inversion efficiency for solar energy utilization. Secondly, extraterrestrial radiation is computed using DEM. At the same time, crop absorbed photosynthetic active radiation is computes based on the fraction of photosynthetically active radiation (fPAR). And the fPAR is computed using normalized difference vegetation index and spectral ratio index resulting from HJ-1 CCD remote sensing images. At last, the inversion APAR and efficiency for solar energy utilization are used to estimate crop net primary productivity. Experiment results indicate that the inversion result of 24 hour actual evapotranspiration is from 4.43mm/d to 8.18mm/d, mean value of which is 6.28mm/d. This value is close to 7. 15mm/d, which is the computed result using Penman-Monteith formula. This indicates our inversion accuracy can meet our application requirements. The mean value of crop NPP 31.02g·m-2·d-1, the highest value is 139.29g·m-2·d-1, the spatial pattern is same to land cover type. These indicate that this study is reasonable and feasible.