<p>The existing credit rating methods usually consider the probability of default (PD) or the loss given default rate (LGD); however, the loan amount is also particularly important to the credit rating. For two loans with the same PD and LGD but with significantly different loan amounts, the degree of economic loss is completely different and, therefore, the credit rating should also be different. If the credit rating is based solely on PD or LGD without taking account of the difference in loan amounts, it is inevitable that the credit rating of two such loans cannot be effectively differentiated. In this study, we propose a credit rating model based on the difference in loan amounts to solve the above rating problem. By dividing customers into different bands according to the difference in loan amounts, and by combining the default pyramid principle with the principle of bell-shaped distribution in each loan amount band, a multi-objective credit rating model is constructed. The objective is to infer a set of optimal thresholds for loan amounts that satisfy both the default pyramid principle and the bell-shaped distribution to achieve a significant difference in loan credit ratings due to different loan amounts. Two datasets of Chinese small enterprises and US peer-to-peer (P2P) personal loans are used to validate the proposed model, and the model proposed can effectively differentiate the credit ratings of two loans with ‘the same default probability and default loss rate, but with significantly different loan amounts’, and outperforms the two comparison models.</p>