Models for speech recognition using deep learning are reviewed for a deeper understanding of the speech recognition process. A key focus of this survey is the recent proliferation of deep learning techniques in speech recognition. Extensive study of deep learning models is presented with a comparative analysis. Different speech recognition tools were presented in paper. Difference evaluation metrics for deep learning-based speech recognition has been illustrated. Comparative analysis for different speech recognition methods was presented. The research finding show that when different combined deep learning model showed reduce word error rate. The research work’s primary contribution is through theoretical and technical details of speech recognition using deep learning. This is accomplished by providing details information about suitable models, tools and feature that can be used to enhance the system. Research finding can be applied to improve deep leaning based speech recognition system.