There are no validated risk scores for predicting coronary heart disease (CHD) in Chinese patients with type 2 diabetes mellitus. This study aimed to validate the UKPDS risk engine and, if indicated, develop CHD risk scores. A total of 7,067 patients without CHD at baseline were analyzed. Data were randomly assigned to a training data set and a test data set. Cox models were used to develop risk scores to predict total CHD in the training data set. Calibration was assessed using the Hosmer-Lemeshow test, and discrimination was examined using the area under the receiver-operating characteristic curve in the test data set. During a median follow-up of 5.40 years, 4.97% of patients (n = 351) developed incident CHD. The UKPDS CHD risk engine overestimated the risk of CHD with suboptimal discrimination, and a new total CHD risk score was developed. The developed total CHD risk score was 0.0267 x age (years) - 0.3536 x sex (1 if female) + 0.4373 x current smoking status (1 if yes) + 0.0403 x duration of diabetes (years) - 0.4808 x Log(10) (estimated glomerular filtration rate [ml/min/1.73 m(2)]) + 0.1232 x Log(10) (1 + spot urinary albumin-creatinine ratio [mg/mmol]) + 0.2644 x non-high-density lipoprotein cholesterol (mmol/L). The 5-year probability of CHD = 1 - 0.9616(EXP(0.9440 x [RISK SCORE - 0.7082])). Predicted CHD probability was not significantly different from observed total CHD probability, and the adjusted area under the receiver-operating characteristic curve was 0.74 during 5 years of follow-up. In conclusion, the UKPDS CHD risk engine overestimated the risk of Chinese patients with type 2 diabetes mellitus and the newly developed total CHD risk score performed well in the test data set. External validations are required in other Chinese populations.