To maintain security， reliability and economy of power system operation， predict credible information in advance is crucial for the power system dispatch and control. Based on research on power system state forecasting， state estimation， and concerned algorithms， a state forecasting and estimation model considering the nodal perturbation power is constructed. In this model， the state estimation results are used as initial value; the quasi-steady process of total variation before and after disturbances are taken into account; and the measurement characteristics of nodal perturbation power due to the randomness of load demand， the self and performing uncertainty of generators scheduling control plan is considered. Weighted least squares and high order correction techniques are employed to forecast and estimate variation trends of system states. The testing results of a real power grid verify the feasibility and effectiveness of the proposed method.