The invention relates to a Kalman filtering-based self-correction cell state-of-charge (SOC) estimation method. The method comprises the following steps that: (1), a state true value and a state estimation value are initialized, wherein the state true value can be determined by an open-circuit voltage method; (2), an SOC calculation formula of the amper-hour integration approach is used as a state equation in the Kalman filter algorithm to calculate the value of Xk+1; (3); the equation is observed to calculate a true value Yk+1; (4), on the basis of the Kalman recursion principle, an SOC value at the current time is corrected continuously and priori estimation error calculation, Kalman gain calculation, correction pre-estimation value calculation, and an optimum estimation error calculation are carried out during the process, so that the estimation value approaches the true value gradually; and (5), according to the certain standard, the Kalman recursion process is ended. According to the invention, the error accumulation effect caused by utilization of the amper-hour integration approach as time goes on can be solved and thus the SOC value can be converged to be one close to the true value soon.