The aim of study was to estimate the current net primary productivity (NPP) of Goksu River Basin (forest, grasslands, bare soil, agriculture) located at the Eastern Mediterranean coast of Turkey using remote sensing and a biogeochemical model. Four elevation zones between 0 to 2500 m were defined and spatial patterns of NPP in those elevation zones were assessed to understand the impacts of topographyon local spatial patterns of productivity. The model results are incoorporated with available topographic information in watershed level. The Carnegie-Ames-Stanford approach (CASA) model approach was used to estimate annual and monthly NPP. This model uses a light-use efficiency (LUE) factor, which is the efficiency of conversion of light energy into dry materials by plant, together with remotely sensed, climate (to express the effects of air temperature and water stress), soil and biotic data and ground measurements. Thus, a comprehensive spatial and temporal data set including temperature, precipitation, solar radiation, soil texture, percent tree cover, land cover type, and normalized difference vegetation index (NDVI) were used in modelling process. Percent tree cover was predicted using multi-temporal LANDSAT images by aggregating tree cover estimates made from high resolution Geo-EYE imagery in a regression tree algorithm. The results indicated several interesting trends between NPP and regional climate gradients. NPP was correlated strongly with solar radiation and precipitation during the growing season suggesting that water limitation as important variable controlling regional patterns of productivity.