Recognizing images of agricultural plant diseases is crucial in the field of intelligent agriculture. A number of cutting-edge machine learning methods, including deep learning and transfer learning, have started to be utilized to identify agricultural diseases in recent years as artificial intelligence technology has evolved. But there are still some significant obstacles to the implementation of these techniques. In particular, machine learning and transfer learning are studied in this work together with current developments in their application to the recognition of agricultural disease images. Transfer learning is preferable given the available agricultural disease data sets, according to analysis and comparison of these two approaches. Automatically detecting plant disease makes crop monitoring easier by allowing for the early detection of disease indications on plant leaves. Fungi, bacteria, and viruses are the primary causes in most plant diseases. Different plant diseases are detected using image processing techniques. This technique includes several processes, including input picture processing, feature extraction, and categorization using different criteria. It uses a range of classification techniques, including K Nearest Neighbour classifiers, fuzzy logic, neural networks, support vector machines, artificial neural networks, and k-means classifiers. Because the effectiveness of the results can change depending on the input data, choosing the optimum classification method can be challenging. This paper focuses on several computer methods as well as various classification algorithm for classifying agricultural diseases.