In the past few years, the prevalence of plant leaf diseases has emerged as a significant issue, necessitating the need for precise investigation and prompt use of deep learning techniques in the categorization of plant diseases. Beans are considered to be one of the most significant plants and seeds that are widely used globally for culinary purposes, whether in their dried or fresh state. Beans are considered to be a very valuable protein source that confers several health advantages. However, the cultivation of beans is sometimes impeded by various illnesses, including angular leaf spot disease and bean rust disease. Therefore, it is essential to establish a precise categorization system for bean leaf diseases in order to effectively address the issue in its initial stages. The objective of this study is to use a deep learning approach for the precise identification and categorization of bean leaf disease. There are several illnesses that are often linked with bean leaves, including angular leaf spot disease and bean rust disease, which have a detrimental impact on bean yield. Therefore, in order to address these issues in their initial stages, a suggested solution involves the use of a deep learning technique. This strategy aims to accurately detect and classify bean leaf diseases by using a publicly available collection of leaf images and employing the MobileNet model, which is implemented using the open source library TensorFlow. The MobileNetV2 architecture was used to train the model under controlled settings, aiming to assess the potential benefits of quicker training times, improved accuracy, and simplified retraining compared to the MobileNet design. The algorithm was evaluated using the bean leaf dataset, and the findings indicate that our approach exhibits a higher level of accuracy in detecting faults.