Objective: To explore the classification of cell types in a breast cancer microscopic image (BCMI), an area-based texture analysis of color space is proposed to classify 3 classes: positive cells, negative cells, and noncancer cell images. Study Design: The BCMI data consisted of 1400, 1200, and 800 images with window sizes 128×128, 192×192, and 256×256, respectively. The texture features based on gray level co-occurrence matrices and fractal dimension are computed from the cell area from 12 color channels and a grayscale image. The optimal vector of features is selected by the sequential forward floating selection (SFFS) method. Moreover, 5 classifiers, namely decision tree, neural network, support vector machine, k-nearest neighbor (KNN), and Naive Bayes are compared and selected for the classification. Results: The result shows that the SFFS method is effective for selecting the best feature vector of BCMI classification. The highest accuracy is 98.88% obtained from the combination of 25 features, window size of 256×256, and KNN classifier. Conclusion: The area-based texture analysis of color spaces can improve the accuracy and increase the ability of BCMI classification that is difficult to classify if a color feature and a spatial feature, notably shape, of cells are similar to each other. © Science Printers and Publishers, Inc.