Reinforcement learning with social payoff promotes cooperation in games. By adding social payoff to Q-learning, agents learn to cooperate for better outcomes. This new approach helps cooperators thrive even when facing strong temptations to defect. The system forms a checkerboard pattern that boosts agents' rewards. The method is robust across different networks and enhances cooperation in spatial games.