Using field research data of 366 family farms in Huang-huai-hai Plain，based on global Moran's I test，the Bayesian spatial Durbin probit model was developed，a spatial weight matrix was constructed，and the optimal model for estimating parameters was selected.Direct effects and spatial spillover effects of characteristic variables on family farms' adoption behavior of green control technique（GCT）were determined by using partial differential method.Results show that GCT adoption behaviors of adjacent family farms have a spatial correlation，and their geographic distribution is spatially clustered.When family farms are within a 2.0 km distance，their GCT adoption behaviors have strong spatial dependence.Education level and risk preference degree，fund status and number of laborers，cognition degree on GCT and the dangers of chemical pesticide use，strength of media publicity，know GCT adopters，and participation in technique training of family farms have significant positive effects on their GCT adoption behavior.The GCT adoption behaviors of family farms are mainly influenced by direct effects of their characteristic variables.However，the spatial spillover effects of neighbor family farms' characteristic variables，especially their participation in technique training，number of laborers，and fund status，cannot be ignored.