Abstract: The classical Weighted Least Squares(WLS) algorithm would encounter the thresholding effect when the noise level was in high level. The Constrained Weighted Least Squares(CWLS) algorithm was applied to multi-station passive localization, the CWLS algorithms based on TDOA(time differences of arrival) and TDOA jointed FDOA(frequency differences of arrival) were proposed. The algorithms first transformed the nonlinear equations into the two pseudo-linear equations, and then added the constraints to the pseudo-linear equations and finally achieved the location. Through simulation results, compared with the classical WLS algorithm and its improved algorithm, the new algorithms extended the operating range with little increase in the computation load. In addition, the new algorithms could still obtain an accurate location when the noises rose beyond the thresholding point. That means the new algorithms are robust to noise and can overcome the thresholding effect effectively.