In the current financial management, there are problems of low efficiency and inaccurate analysis data, which have brought great inconvenience to the financial management, thus greatly hindering the development of enterprises. Machine learning algorithm and computer vision technology have many applications at present, and also have applications in finance, but no one has applied them to the model of financial accounting image management system. Based on this, this paper applied machine learning algorithm and computer vision technology to the model of financial accounting image management system. This paper first introduced the application of the reimbursement system management method, and then analyzed the the current issues in financial management. Next, it introduced the machine learning algorithm and computer vision technology, and designed the model of the financial reimbursement image management system. Finally, the application effect of the model of financial reimbursement image management system was analyzed. This paper compared and analyzed the model of the general financial reimbursement image management system and the model of the financial reimbursement image management system proposed in this paper. It was concluded that the model of financial accounting image management system oriented to machine learning algorithm and computer vision technology has better operation convenience compared with the model of general financial accounting image management system, which can greatly improve the accuracy of analyzing financial accounting data and the risk prevention and control ability of the financial management department. The effect of financial management also improved, and the satisfaction of employees with profit distribution management increased by 29%. It can be applied to financial management in the future to improve the effect of financial management.