Spiking Neural Networks (SNNs) are garnering growing interest and research attention in both academia and industry, owing to their striking similarity to the neural mechanisms of biological brains. However, the application of SNNs in the realm of robot control still presents certain challenges. In this work, a bioinspired autonomous learning algorithm for the mobile robot is proposed, which is grounded in reward-modulated spike-timing-dependent plasticity (R-STDP). A novel reward generation mechanism is employed to generate reward signals that guide the learning and decision-making processes. The robots validate the efficacy of the algorithm through an obstacle avoidance task in both simulated and real-world environments, demonstrating their ability to successfully navigate around obstacles after undergoing autonomous learning. Experimental results underscore the algorithm’s effectiveness, and the mobile robot equipped with this algorithm learning capabilities in classic task scenarios. Experimental results demonstrate the algorithm’s effectiveness, and the mobile robot equipped with this algorithm exhibits impressive autonomous learning capability. This work offers an alternative approach to designing mobile robots endowed with autonomous learning abilities and biological characteristics, suitable for deployment in classical task scenarios.