Induction motors operate in difficult environments in the industry. Monitoring their performance in such circumstances is significant, as it can provide a reliable operation system to secure the production line. Recently, Artificial Intelligence techniques (AI) were applied to the condition monitoring and fault diagnosis systems in order to build an efficient classification model. This paper focuses on developing a new hybrid diagnosis model for fault classification. The development of this model provides a novel technique for the diagnosis of single and multiple induction motor faults. The aim is to find a new alternative source to extract automatic features from the motor parameters. Three deep learning networks including Visual Geometry Group 19 model (VGG-19), Residual Network 50 model (ReseNet-50), and EfficientNet-B0 model (EffieNet-B0) were applied to pre-train the suggested model. The use of these networks can also allow the attributes to be automatically extracted and associated with the decision-making part. The model's performance was assessed by calculating some evaluation metrics, such as the confusion matrix, accuracy, precision, recall, and F1 Score. The evaluation of the proposed model was achieved by applying different types of motor data including stator current data and motor vibration data. In addition, Convolutional Neural Networks (CNNs) were applied as an image processing method to achieve the model features. The experimental results proved the robustness and capability of the proposed model for fault classification by combining the suggested networks. The suggested hybrid model achieved a high classification accuracy.