Image contrast of fiber gyroscope fiber defects was low, at the same time object and background didn’t discriminate easily. Fuzzy image enhancement processing was studied and part low gray level was set zero, which leads to original image information loss using the Pal fuzzy algorithm. Combine minimum bias-normal distribution index and modified fuzzy algorithm was put forward and spacing point is counted by minimum bias-normal distribution index and then automatically affirms sub region and nonlinearity transformation was used in inverse process. This algorithm can engender different enhancement effect in different image region, which enhance contrast ratio in object and background, outstanding defect character. The algorithm was used for fiber gyroscope fiber defect infrared image processing. Experiment results show that the algorithm is effectiveness.