Abstract: The surface corrosion morphology of 2xxx aluminum alloy specimens pre-corroded with 3.5% mass fraction NaCl solution for different times was analyzed by digital image processing technique. The three-dimensional corroded morphology was described by Differential Box-Counting method. The relationships between fractal dimension, corrosion time and fatigue life were analyzed, and the fatigue life prediction model based on fractal dimension was established and verified by experiments. The results show that the relationship between the fractal dimension and corrosion time was a power function, while the fractal dimension was linear with the fatigue life. The prediction results of the established life prediction model for low cycle fatigue lives were relatively conservative, and the relative error with the test result was within 25%. The prediction results for high cycle fatigue lives were relatively accurate, and the relative error of the test result was reduced to less than 15%. The fatigue limit calculated by life prediction model was close to the test value, and the relative error of less than 5%.