The present invention relates to an adaptive correction method for magnetic resonance diffusion weighted imaging, comprising the following steps: step 1, repeating acquisition of diffusion weighted images N times with the same scanning parameters, N≥3; step 2, based on the original image or compressed image point by point Construct a correlation matrix; step 3, perform principal component analysis on the correlation matrix after smoothing and filtering, and obtain the eigenvector corresponding to the largest eigenvalue of each correlation matrix; step 4, calculate the weight according to the eigenvector; step 5, according to Weighting performs weighted synthesis on the original image to obtain a corrected diffusion-weighted image. On the basis of multiple acquisition average technology, the present invention adopts principal component analysis method to adaptively detect and correct data from redundant data, suppress motion artifacts, radio frequency ignition artifacts, etc., and improve image quality; no need to increase Hardware device, and the image quality is better than multiple acquisition direct averaging technology.