The invention discloses a weight function least square state estimation method based on residual normalization, and belongs to the field of electric power system dispatching automation. The method comprises the steps that SCADA data, collected by a data terminal, of any time section, a network structure and parameter information are input firstly through a computer and a program and initialized; a node admittance matrix of a network is calculated, a zero-injection equality constraint equation is formed, and then a corresponding residual, a jacobian matrix and a weight function are calculated by comprehensively considering a voltage amplitude value measuring equation, an injection power measuring equation and a branch circuit power measuring equation and taking the node voltage amplitude value and phase angle as state variables; finally, the state variables are updated, convergence judgment is conducted, and the state estimation of a power grid is achieved. By means of the weight function least square state estimation method based on residual normalization, bad lever measurement can be effectively restrained, high robustness and good convergence are achieved, the computational efficiency is very high, and a good engineering application prospect is achieved.