Based on the concept of normal measuring point， evaluation function of measuring point is constructed to judge whether a measuring point is normal or exceptional. Then， a new model of state estimation is proposed which takes the maximum of normal rate of measuring point as the objective. The model can consider power flow constraints and other physical constraints to make the estimation results close to the actual scenario. The model can be successfully solved by modern interior point method which has short CPU time and good convergence property. Comparing with other state estimation methods， the proposed method has the following advantages: higher normal rate of measuring point， capability of automatically identifying bad data， stronger anti-error characteristics， and capability of obtaining solutions suits practical scenario better because of consideration of power flow constraints and other physics constraints in the solution process. At the same time， the method does not need to perform additional observability analysis， bad data detection or subjective weighting factors setting， thus substantial reduction in maintenance efforts can be achieved.