In the current corporate financial management, the traditional financial budget control method relies on static rules and historical data, which is difficult to cope with the complex and changing market environment and internal operation fluctuations, resulting in unreasonable budget allocation, low resource utilization, insufficient dynamic adjustment capabilities and other problems. It employs reinforcement learning technology to create a dynamic financial budget intelligent control mechanism in a bid to improve the flexibility and efficiency of the budget allocation process. First, an environmental model of financial budget control is constructed from the enterprise financial status, market dynamics, and operation needs for the state space, with the budget allocation plan for the action space, and financial performance indicators for the reward signal. The DQN intelligent control model exists and trains in order to optimize the budget allocation policy through trial and error and a reward mechanism. Finally, it allows flexibility in adjusting the budget allocation plan with real-time data. The mechanism raised the budget utilization rate to 91%. The greatest net profit achieved in the financial performance indicator is 350,000 yuan. The dynamic financial budget intelligent control mechanism which is based on reinforcement learning is capable of enhancing the flexibility and efficiency levels of corporate financial management and provides an intelligent solution to control budgets in an intricate environment.