An automatic exposure optimization method for mitigating overexposure, slow convergence speed, and instability in HDMI digital microscopy cameras due to the inconsistent reflection levels of different targets during observation is proposed. Compared with a traditional automatic exposure control method, the proposed method facilitates exposure control based on the region of interest, metering, and brightness statistics on any area in RAW images, thus expanding the dynamic range of brightness statistics and facilitating exposure control for different targets. In addition, the method improves the accuracy of brightness statistics and mitigates overexposure caused by extremely bright or dark targets. A variable exposure adjustment step size is set for the dynamic adjustment of exposure step size according to the current brightness level, achieving coarse and fine adjustment of exposure to balance exposure convergence speed and exposure accuracy. Finally, experiments were performed using actual cameras. The proposed method reduced the exposure deviation by half compared with the traditional automatic exposure method. The exposure convergence time was shortened by more than half, and the fastest exposure convergence was achieved within four frames of an image.