The synthetic aperture passive localization system generally compensates for the second-order phase term of the received signal with the Taylor series of the range history and then uses the focusing result of the compensated signal to obtain the position of the emitter. However, the existence of a higher-order residual phase causes the mismatch of reference function, leading to the bias of localization results. To solve the problem, this paper proposes a slant range expansion method based on an orthogonal basis. The optimal expansion of the range history is obtained by constructing a set of orthogonal bases in the space composed of quadratic polynomials so that the residual phase after integration is minimized. The proposed method can effectively mitigate the localization bias caused by the model approximation of a synthetic aperture localization system. Simulations and Monte Carlo tests show that the proposed method outperforms the traditional synthetic aperture localization method.