In bearings-only localization,the target position is obtained from the point of intersection of bearing lines emanating from two or more observer positions.However,the measurement equation is nonlinear,and the actual noise is difficult to be modeled accurately.To solve the problem,a direction cross target location filtering equation is constructed,and an algorithm combining smooth variable structure with Kalman filtering is proposed based on switching gain strategy.On this basis,the residual adaptive algorithm is used to further improve the accuracy of the combined filter,which effectively improves the accuracy of target locating and the robustness of the algorithm in the case of large disturbances,noise uncertainty and so on.The result of simulation experiment shows that:1) The locating accuracy of the combined filtering algorithm is higher than that of the unscented Kalman filtering algorithm and the adaptive iterative unscented Kalman filtering algorithm in the case of large disturbance;and 2) In the case of non-Gaussian noise distribution,the unscented Kalman filtering algorithm and the adaptive iterative unscented Kalman filtering algorithm show a divergent trend,while the combined filtering algorithm can still converge stably.