Leukemia, a severe hematologic malignancy, requires rapid and accurate detection for effective treatment. This study presents an automated classification system leveraging image processing and machine learning. Preprocessing involves grayscale conversion, Gaussian blurring, and thresholding for binary segmentation. Feature extraction includes identifying cell contours, calculating total cell count, and assessing average cell area. The extracted features are classified using Support Vector Machine (98% accuracy), Random Forest (99% accuracy), and Gradient Boosting Classifier (99% accuracy). Performance evaluation using a confusion matrix and accuracy score demonstrates the system’s effectiveness in distinguishing normal and leukemia samples. Automated image-based classification enhances diagnostic accuracy and reduces reliance on manual examination, with future work aiming to integrate deep learning for improved robustness and real-time application.