Sugarcane is a special crop that developed from a wild grass species to a major crop utilized for food and energy as a result of human activities. Due to its substantial contribution to the national finances and employment of more than a million people, mostly in the production of sugar and renewable bioenergy, it is a major crop that requires careful monitoring. Crop losses for farmers, both in terms of quantity and quality is caused by diseases in crop. In order to increase agricultural output and performance, scientists are now focused on early disease detection and identification. Additionally, agriculture makes substantial use of artificial intelligence. This study attempts to classify diseases in sugarcane using transfer learning and deep learning algorithms. In this context, first a detailed review on the recent contribution in sugarcane disease detection is discussed. Then, using leaf images as data, deep learning techniques are implemented utilizing transfer learning to build models that can identify sugarcane diseases. Metrics like confusion matrix, accuracy, precision, recall, and f1-score are used to evaluate the performance of various deep learning models. Different deep learning models such as Convolutional Neural Network (CNN), Visual Geometry Group 16(VGG16), Visual Geometry Group 19 (VGG19), Alex Krizhevsky’s Network (AlexNet), Inception convolutional neural network Version 3 (Inception V3) and Densely Connected Convolutional Network with 201 layers (DenseNet201) were trained and results show that DenseNet201 outperforms compared to others with 94.14 % accuracy. This study not only shows how transfer learning can improve sugarcane disease classification, but it also verifies that deep learning models can function well in classification-related tasks.