An early and precise diagnosis of prostate cancer (PCa) is essential to enhance treatment effectiveness. The detection of prostate cancer typically relies on human expertise and effort, which may sometimes lead to mistakes that could put a patient’s life at risk. Therefore, the use of artificial intelligence (AI) using deep learning (DL) algorithms is proposed to assist experts in reducing diagnostic errors in the identification of prostate cancer. This research investigates how deep learning model, particularly MobileNet, can be used for automatic prostate cancer detection. MobileNet is known for its efficiency on mobile and embedded devices and offers a lightweight architecture that facilitates rapid image analysis with minimal computational resources. The proposed model was trained and tested using a magnetic resonance imaging (MRI) dataset containing both benign and malignant prostate tissue. The initial stages included data augmentation to enhance the diversity and quality of the training data. Performance metrics were used to evaluate the model during the training process, yielding the following outstanding results: accuracy (94%), precision (92%), recall (100%), and F1 score (96%) in identifying and categorizing prostate cancer. In summary, the research demonstrates how the deep learning model, specifically MobileNet, can enhance the effectiveness of detecting and classifying prostate cancer.