The asymptotic single risk factor approach (ASRF) approach is a simplified framework for determining regulatory capital charges for credit risk and has become an integral part of how credit risk capital requirements are to be determined under the Basel II. Within this approach, a key regulatory parameter is the average asset return correlation. In this paper, I examine the empirical relationship between the asset return correlation, firm probability of default and firm asset size in five models. According to previously theoretical background, I separate these models into two parts. First part I called Market Correlation Model, which conclude Basel and KMV’s factor Model, and the results in these models advocate the explanation for a higher asset return correlation of large firms may be that they are better diversified than small firms. Second part I called Asst Value Correlation Model, which conclude one factor model; default correlation model and structure model, and the results in these models advocate the issue of “financial accelerator” which means macroeconomics shocks should have a stronger impact on small firms and we should observe a higher asset return correlation. Especially in structure model which I devised, it not only concludes the advantage of Market Correlation Model but also combines the default point in Asset value correlation model. Furthermore, I prove the one factor model and default correlation model under some conditions are the same. In our empirical results, we can find that the range of asset return correlation in Basel Model is the largest range number than others and I also find the volatility of probability default has a strong impact on asset return correlation.