The accurate and efficient classification of leukemia images is crucial for early diagnosis and effective treatment planning. Traditional methods often face challenges in handling the complexity and variability of medical images. To address these challenges, we propose a novel approach that leverages the Gray Level Co-occurrence Matrix (GLCM) and statistical feature-based segmentation techniques. In this paper, we present a comprehensive framework for the automated classification of leukemia images using advanced image processing techniques. The methodology involves six key stages: input of leukemia images, preprocessing to enhance image quality, segmentation to isolate relevant features, feature extraction using texture analysis, classification using multiple distance metrics Euclidean, Manhattan, Canberra, and Chebyshev, and performance evaluation. Our results demonstrate significant improvements in classification accuracy, sensitivity, specificity, and error rates across various metrics and feature sets. For instance, using the Chebyshev distance, we achieved an average accuracy of 82.69%, sensitivity of 85.95%, and specificity of 82.77%. The Canberra distance provided optimal performance with 65 features, yielding an accuracy of 85.18%, sensitivity of 86.39%, and specificity of 86.31%. These findings underscore the efficacy of our approach in distinguishing between healthy and leukemic cells, thereby contributing to early diagnosis and effective treatment planning for leukemia.