This paper made a comprehensive discussion about the spatial differentiation, evolution trend, correlation and driving factors of the regional agricultural mechanization level by using TOPSIS method, trend surface analysis, ESDA and GWR model. The results show that Hebei county’s agricultural mechanization level presents obvious spatial directivity and topographical distribution differences in 2013. The agricultural mechanization level develops better in central southeast plain, followed by the northwest plateau, and hilly region relatively worse. There exists a significant spatial autocorrelation characteristic and regional convergence phenomenon. The southern area of Hebei major in HH type, and northern area major in LL type. GWR shows great superiority in explaining the spatial non-stationary of elements, and reveals both positive and negative correlations between farmland scale and plant structure (expect for terrain), which is different from OLS result that all the factors are positive. In the future the government can consider such measures as enhancing the cultivated land scale and proportion of planting structure in the plateau area, considering other factors in the plain area to promote county’s agricultural mechanization level.