We introduce FinMR, a novel multimodal benchmark designed to evaluate the reasoning capabilities of multimodal LLMs in financial problem-solving. FinMR features 3,200 college-level question-answer pairs, including 1,049 focused on financial math and 2,151 on financial expertise. It integrates both textual and visual content, such as financial tables and stock price trends. All answers include expert-annotated explanations, enabling detailed analysis of reasoning errors. To enhance financial reasoning, we propose Error Feedback Learning (EFL), which uses negative examples and feedback for iterative improvement. Through evaluations of open-source and closed-source models, we demonstrate that MLLMs outperform LLMs and that EFL is more effective than CoT prompting. Our error analysis highlights key challenges in image recognition, question understanding, and formula application, providing insights for future research. FinMR establishes a robust foundation for advancing financial reasoning capabilities and developing more effective multimodal reasoning techniques.