Abstract: In the development of critical heat flux (CHF) correlation, the mixing factor β (i.e., the turbulent thermal diffusivity) is a characteristic parameter representing the cross-mixing effect between sub-channels. At present, the mixing factor obtained through the cross-mixing experiment is essentially the macroscopic equivalent quantity of averaging the cross-mixing effect (such as turbulent cross-mixing, flow sweeping, etc.). In fact, the mixing effect caused by the grid will decrease along the downstream, which will lead to a certain deviation in the local parameters calculated by the average mixing factor. The influence of the deviation on the DNBR limit value of the CHF correlation and the safety margin is not clear. Thus, the CHF correlation matching with the sub-channel analysis program FLICA-ⅢF was developed by using the minimum DNBR point method. Owen criterion was used to determine the DNBR limit. After achieving the development methods of CHF correlation, the prediction effect, DNBR limits and actual maximum power of correlations of different mixing factors were compared and analyzed. The results indicate that in the range of experimental conditions, the local pressure of minimum DNBR point does not change with the change of mixing factor, the local mass flow rate increases with mixing factor, and the local quality decreases with the increase of mixing factor. The deviation of mixing factor will increase the DNBR limit of its correlation, but there is no linear relationship between the two. Besides, in actual conditions, the maximum power calculated by the correlations of different mixing factors is different. While the maximum power value of the precise mixing factor is also the largest. In conclusion, the more precise mixing factor determined by experiment has higher accuracy and economy in fitting correlation. Therefore, it is valuable and meaningful to consider the fine subchannel of grid mixing effect.