The goal of this research is to find a solution of multiclass classification of medical images using Deep Learning. An analysis of the effectiveness of the most common pre-trained models was carried out. Transfer learning was chosen as the methodology. The article presents a description of the VGG-19, ResNet-50, InceptionV3 models and their comparative characteristics. The dataset for this study consisted of chest X-Ray images sourced from prominent healthcare organizations in the field, including the Society for Imaging Informatics in Medicine. Additionally, the dataset received contributions from the Foundation for the Promotion of Health and Biomedical Research of Valencia Region, the Valencian Region Medical ImageBank, and the Radiological Society of North America. Notably, the ResNet-50 based model demonstrated the highest performance, achieving an impressive accuracy of 82.74%. The corresponding Tables reflect the outcomes for each model. The practical significance of the study is that the results obtained can be used to speed up the process of diagnosing medical diseases.