This paper first analyzes the status quo of regional economic law coordinated development and draws out the mechanism and characteristics of regional economic law legal imbalance under the perspective of game theory. Then, on the basis of the proximal strategy optimization algorithm, the generative adversarial network is introduced, and according to the auxiliary role of the subjective metric constituted by the generator in the generative adversarial network as the dominance function to provide the real metric, the GAN-PPO algorithm based on the generative adversarial network is constructed to improve the GAN-PPO algorithm. Finally, after evaluating the performance of the improved PPO algorithm, the mechanism and degree of imbalance in regional economic law laws are studied using the GAN-PPO algorithm. The results show that the GAN-PPO algorithm reaches the peak gain at 1,250,000 steps, the peak value of the average reward obtained is 885, and the speed of convergence is improved by 0.42 in the sensitive environment, which realizes the acceleration of the algorithm. In the regional economic law imbalance, the degree of regional economic imbalance is between [0.1,0.48], the degree of industrial imbalance is between [0.08,0.43], the degree of capacity imbalance is between [0.10,0.45], and the degree of information imbalance is between [0.05,0.38], and the regional economic law legal imbalance is obvious, and it needs to be reasonably regulated.