Machine learning (ML) is one of the fields of artificial intelligence that offers algorithms topredict from samples the effective detection of skin lesions caused by skin cancer. This paper presents thepreliminary identification of skin lesions using optimized algorithms for texture feature extraction byGLCM and feature-based learning (LightGBM, SVM and HAAR Cascade) as an initial stage for adiagnostic tool. The HAM10000 skin lesion image set, Python programming language and open sourcebased libraries are used to process the images, extract the features and train the learning models, determinethe performance and hit rate of the models. Based on the results obtained, the LightGBM classifier requiredthe shortest learning time, reduced CPU usage and 90 % accuracy rate