This paper proposes a feature selection algorithm based on particle swarm optimization (PSO) combined with machine learning methods to improve the accuracy of breast cancer diagnosis. The algorithm was tested on three breast cancer datasets: Coimbra (CD), Wisconsin Diagnostic Breast Cancer (WDBC), and Wisconsin Prognostic Breast Cancer (WPBC). In the Coimbra dataset, the accuracy increased from 87% to 91% with PSO feature selection, reducing the number of features from 9 to 4. In the WDBC dataset, accuracy improved from 99% to 100% with PSO, reducing features from 30 to 19. In the WPBC dataset, accuracy increased from 94% to 96%, reducing features from 33 to 17. The results show that PSO feature selection improves classification accuracy while reducing the number of features. The study also compares the performance of various machine learning algorithms with and without PSO feature selection, demonstrating that PSO enhances accuracy and reduces computational costs. The proposed method is effective in selecting the most relevant features for breast cancer diagnosis, leading to more accurate and efficient classification. The research highlights the benefits of using PSO in feature selection for medical diagnosis, particularly in reducing data complexity and improving diagnostic accuracy.This paper proposes a feature selection algorithm based on particle swarm optimization (PSO) combined with machine learning methods to improve the accuracy of breast cancer diagnosis. The algorithm was tested on three breast cancer datasets: Coimbra (CD), Wisconsin Diagnostic Breast Cancer (WDBC), and Wisconsin Prognostic Breast Cancer (WPBC). In the Coimbra dataset, the accuracy increased from 87% to 91% with PSO feature selection, reducing the number of features from 9 to 4. In the WDBC dataset, accuracy improved from 99% to 100% with PSO, reducing features from 30 to 19. In the WPBC dataset, accuracy increased from 94% to 96%, reducing features from 33 to 17. The results show that PSO feature selection improves classification accuracy while reducing the number of features. The study also compares the performance of various machine learning algorithms with and without PSO feature selection, demonstrating that PSO enhances accuracy and reduces computational costs. The proposed method is effective in selecting the most relevant features for breast cancer diagnosis, leading to more accurate and efficient classification. The research highlights the benefits of using PSO in feature selection for medical diagnosis, particularly in reducing data complexity and improving diagnostic accuracy.