As an emerging powerful tool to provide structural information of tissue specimens label-freely, Mueller matrix (MM) polarimetry has garnered extensive attention in biomedical studies and pathological diagnosis. However, for the commonly used constant-step rotating MM polarimetric system, beam drift induced by the rotation of polarization elements can lead to distortions in measurement results, severely affecting MM imaging accuracy. Here, based on our previous study, we propose an optimized self-registration method to mitigate the pseudo-depolarization effects introduced by image artifacts in constant-step rotating MM polarimetry. By addressing the prevalent issue of beam drift and image distortions in such polarimetric imaging systems, the effectiveness of the proposed method is experimentally validated using tissue samples. The results demonstrate a significant enhancement in the accuracy of depolarization parameter estimation after applying the optimized self-registration method. Furthermore, the method enhances the coarseness and contrast of MM-derived parameters images, thereby bolstering their capacity to characterize tissue structures. The optimized self-registration method proposed in this study can provide an innovative approach for quantitative tissue polarimetry based on constant-step rotating MM measurement, and contribute to the advancement of polarimetric imaging technology in biomedical applications.