A statistic method,statistics of extreme values(SEV),was described in detail,which can estimate the size of maximum inclusion in steel.The characteristic size of the maximum inclusion in a high clean bearing steel(GCr15)was evaluated by this method,and the morphology and composition of large inclusions found were analyzed by scanning electron microscopy(SEM).When standard inspection area(S0)is 280 mm~2,the characteristic size of the biggest inclusion found in 30 standard inspection area is 23.93μm,and it has a 99.9% probability of the characteristic size of maximum inclusion predicted being no larger than 36.85μm in the experimental steel.SEM result shows that large inclusions found are mainly composed of CaS,calcium-aluminate and MgO.Compositing widely exists in large inclusions in high clean bearing steel.Compared with traditional evaluation method,SEV method mainly focuses on inclusion size,and the estimation result is not affected by inclusion types.SEV method is suitable for the inclusion evaluation of high clean bearing steel. A statistic method, statistics of extreme values ​​(SEV), was described in detail, which can estimate the size of maximum inclusion in steel. The characteristic size of the maximum inclusion in a high clean bearing steel (GCr15) was evaluated by this method, and the morphology and composition of large inclusions found were analyzed by scanning electron microscopy (SEM) .When the standard inspection area (S0) is 280 mm ~ 2, the characteristic size of the largest inclusion found in 30 standard inspection area is 23.93 μm, and it has a 99.9% probability of the characteristic size of maximum inclusion predicted being no larger than 36.85 μm in the experimental steel. SEM result shows that large inclusions found are mainly composed of CaS, calcium-aluminate and MgO. Compositing widely exists in large inclusions in high clean bearing steel. Compared with the traditional evaluation method, SEV method mainly focuses on inclusion size, and the estimation result is not affected by inclusion types. SEV method is suitable for the inclusion evaluation of high clean bearing steel.