Abstract: To solve the problem that traditional long baseline (LBL) positioning system ignored the calibration errors or underwater beacon array errors, this paper proposes a method combining calibration errors of undersea beacons and distance measurement errors together. This method uses a priori error model of observations to build up the error adjusting equations, with which the target position can be optimized. This method is used to process actual shallow water data, the position trajectory is continuous and smooth. Compared with traditional method, this method has the estimation more approaching to GPS outputs. Theory analysis and experiment results show that using this method can achieve more optimized estimation than traditional method. This method can meet the requirement for high-precision long baseline acoustic positioning systems.