Abstract: Greenhouse temperature control is very complex because it is affected by several fators such as solar radiation, wind and external temperature. In this paper specialized control scheme of proportional integral(PI) control with feedforward was developed to solve these problems. In practical use, the suitable parameters are the key to obtain satisfactory control result. Based on radial basis function neural networks the model was developed to predict the inside air temperature in greenhouse. In the model the outside air temperature, solar radiation, inside relative humility and status of greenhouse were addressed as inputs. Based on this model, the control parameters can be determined quickly.