Mpox is a viral disease that has affected several regions of the world, and early detection is crucial for controlling its spread. However, the visual identification of symptoms associated with the disease in medical images or photographs can be challenging, especially in scenarios of high variability in symptom presentation. Computer vision and machine learning present themselves as effective solutions to address this challenge, enabling automatic image classification and aiding decision making in public health management. The applied methodology was based on 5 phases: Data acquisition; data preprocessing; feature extraction (HOG, Histogram, SURF and LBP); Model building and tuning (Machine Learning (Decision tree and SVM), Deep Learning (EfficientNetB7, NASNet, Vision Transformer and VGG19) and Measurement (Accuracy, Precision, Recall, F1-Score and Validation with images). The superior results were obtained with the NASNet model, with the metrics of accuracy, precision, recall (sensitivity) and F1-score, with the values of 0.994231, 0.994238, 0.994231 and 0.994231, respectively. In conclusion, the results demonstrate the high capacity of deep learning models to extract different complex characteristics automatically and efficiently, which highlights them as sufficiently robust tools to be considered in clinical applications.