Compared with single-mode fibers (SMFs), multimode fibers (MMFs) can support a much larger number of guided modes, offering the attractive advantage of high-capacity information and image transportation. We propose a physics-informed deep learning framework for efficient wavelength-multiplexed image transmission through multimode fiber. Experimental results demonstrate its potential for preserving high-fidelity information transfer while ensuring robustness and high resolution in multimode fiber. Multimode fibers with high information capacity and ultra-thin diameter offer new possibilities for non-invasive endoscopy and remote high-speed secure communication. To overcome this challenge we present a deep learning approach that generalizes over mechanical perturbations.