In this paper, we introduce a new algorithm, fiDrizzle-MU, to coadd multiple exposures with multiplicative updates in each iteration instead of the difference correction terms of the preceding version. We find that multiplicative update mechanisms demonstrate superior performance in decorrelating adjacent pixels compared to additive approaches, reducing noise complexity in the final stacked images. After applying fiDrizzle-MU to the JWST-NIRCam F277W band data, we obtain a comprehensive reconstruction of a potential gravitational lensing candidate substantially blurred by the JWST pipeline’s resampling process.