This article studies the obstacle avoidance algorithm for intelligent agents based on fuzzy neural networks and implements the algorithm. Fuzzy neural networks combine the advantages of fuzzy logic and neural networks, and can handle uncertainty and nonlinear problems, making them suitable for obstacle avoidance scenarios of intelligent agents. Through experimental verification, the algorithm has shown good obstacle avoidance performance in various scenarios, improving the adaptability and robustness of the intelligent agent. The research results indicate that the obstacle avoidance algorithm based on fuzzy neural networks has practical application value, providing new solutions for path planning and obstacle avoidance problems of intelligent agents.