To solve the imaging distortion problem in a nonuniform temperature field for visual measurement, this paper studies the Runge-Kutta ray tracing algorithm. First, the light ray emitted by the spatial light point will be abstracted by the proposed method. Then, the ray trajectory in a 3D nonuniform temperature gradient environment will be reproduced, and the ray distortion can be corrected. Finally, the distortion is compared with the actual image distortion obtained through the corner detection method. The experimental results show that the correction method for the distortion image caused by nonuniform temperature gradient heat source interference is feasible by using the Runge-Kutta ray tracing algorithm, which can realize the error compensation for image distortion and obtain an accurate corrected image.