Abstract: For the most common fuzzy operation in digital image tampering, a digital image forensics algorithm based on the correlation detection is proposed. The algorithm assumes that linear correlation of pixel and neighboring pixels of the original image after fuzzy operation and the image has been tampered is enhanced, which uses the shortest distance clustering method to detect linear correlation of pixels, and reveals the trace of tampering, thus achieves forensics of digital image tampering. Through a lot of experiments, the results show that, the algorithm can effectively detect fuzzy operation trace of the tampered image, at the same time, which can be able to accurately locate the tampered.