Wireless sensor networks (WSNs) are gradually invading our daily lives, offering us new services every day. They can be found in applications that affect us more and more. First used to monitor the environment and urban areas, they then provided support for first-aid and military surveillance activities. Now they are appearing in applications even closer to home to improve our lifestyle, such as guiding us to available parking spaces or informing us about air quality. The wide range of applications for wireless sensor networks has prompted several researchers to work towards a WSN with lower deployment costs while maximizing network lifetime and coverage. In this paper, an optimization approach-based Q-learning algorithm for optimal coverage of heterogeneous sensor networks is proposed. The findings of the simulation prove that the proposed approach maintains network coverage while using the minimum amount of energy, compared with other approaches.