<p>Breast cancer is a serious global health challenge that poses a constant threat to women's lives. Accurate histopathological imaging in breast screening can significantly improve the treatment success rate. In this study, we propose an enhanced sine cosine algorithm (AGSCA) based on an adaptive weight strategy (AW) and globally adaptable hunting strategy (GAH) to obtain optimal thresholds, thereby providing more precise breast cancer pathology images. The AW strategy enhances the algorithm's convergence speed and its ability to escape local optima, while the GAH strategy maintains algorithmic diversity and improves exploration capability. To validate the performance of the proposed algorithm in achieving optimal thresholds, we conduct a series of comparative experiments at 30, 50, and 100 dimensions using the IEEE CEC 2017 test suite, which demonstrated the superiority of AGSCA. Furthermore, by integrating this algorithm with Kapur's entropy and non-local means two-dimensional histogram, we propose a multi-threshold image segmentation model (AGSCA-MTIS). This model is applied in segmentation experiments at six threshold levels on nine representative clinical breast cancer images. The experimental results, measured by the feature similarity index, peak signal-to-noise ratio, and structural similarity index, confirm that AGSCA-MTIS is more effective in handling image details, with more pronounced features in lesion areas and higher overall image quality. Consequently, this study presents a competitive and efficient tool for breast cancer diagnosis, with significant potential for application.</p>