To address errors in transmittance estimation， an underwater polarization image restoration method based on transmittance-limiting correction is proposed. Initially， a coarse correction of the transmittance parameters is applied to suppress noise， thereby reducing interference in target imaging. Subsequently， a fine adjustment of the transmittance parameters is performed to enhance contrast and improve estimation accuracy. The corrected transmittance is then integrated into the underwater transmission model to produce the restored image. An independently developed underwater active illumination imaging experimental device was employed to conduct restoration experiments on various targets in waters of differing turbidity levels. Experimental results demonstrate that the proposed method improves information entropy and contrast indices by averages of 33.89% and 69.01%， respectively， compared to the original images. Additionally， it outperforms other algorithms with respect to average gradient， signal-to-noise ratio， and information entropy. Target details were significantly recovered in turbid waters of varying concentrations. Under active illumination conditions， effective restoration of underwater polarization images necessitates accurate acquisition of the backscattered degree of polarization and the light intensity at infinity. Rough estimations of these parameters may lead to transmittance errors， resulting in considerable noise and pixel saturation in the restored images.