Net primary productivity (NPP) in the coastal salt marshes of the Liaohe River Delta was estimated using Carnegie-Ames-Stanford Approach (CASA) model with remote sensing, the meteorological and field observation data. The spatial difference of NPP and its causes were analyzed. The results showed the following: (1) the spatial differentiation of NPP in the study area was significant with the variation coefficient 43.71%. The NPP within the same vegetation type also changed significantly. (2) N, P, and K limitation of NPP occurred in the form of co-limitation, and it took the form of a parabola with the increase of soil conductivity. The sensitivity was in the order of K > P > N, while the degree of nutrient limitation was K > N > P. (3) The influence of soil moisture and salinity on NPP also had the form of a parabola. Compared with soil moisture, the sensitivity of NPP to soil conductivity and influence degree of soil conductivity on NPP were greater. Meanwhile, the higher soil moisture, the lower sensitivity of NPP to soil conductivity. Our findings highlight the importance of the optimal combination of edaphic factors, which is significance of the management and conservation of wetland vegetation.