BACKGROUND Diabetes mellitus (DM) and coronary heart disease (CHD) are in a complex medical status which are closely associated and generally coexist. Many relationships of DM & CHD are well-known but their precise nature still remains unknown. OBJECTIVE The present report aims to derive an appropriate probabilistic model between DM & CHD, and also to identify their risk factors, based on a data set of 366 African Americans in rural Virginia, USA. METHODS Both the positive responses glycosylated hemoglobin and total cholesterol are identified as non-constant variance. Thus, they may be modeled by joint log-normal or gamma models, where both the mean and variance are modeled by using two interlinked models for the mean and the variance, based on the observed data and gamma deviance. RESULTS Statistical significant causal factors, namely, total cholesterol (chol) (P=0.002), hdl-c (P=0.082), interaction effect of tchol & hdl-c (P=0.032), stabilized glucose (stb.gl) (P=0.000), postprandial time when labs were drawn (tim.pn) (P=0.024), interaction effect of stb.gl & tim.pn (P=0.008), age (P=2e-7), height (P=0.029), interaction effect of age & height (P=0.003), waist (P=0.014), location (P=0.016), sex (P=0.007), frame of study (P=2e-4), have been identified as the determinants of DM (based on glycosylated hemoglobin (HbA1c)). Identified statistical significant factors for CHD (based on tchol) are interaction effect of hdl-c and ratio of tchol & hdl-c (P=0.000), HbA1c (P=0.045), tim.pn (P=0.020), interaction effect of HbA1c & tim.pn (P=0.019), stb.gl (P=0.016), age (P=0.037), location (P=0.070), sex (P=0.005), height (P=0.002), interaction effect of height and tim.pn (P=0.007), first diastolic blood pressure (bp.1d) (P=2e-4), interaction effect of bp.1d & first systolic blood pressure (bp.1s) (P=0.084). CONCLUSIONS It has been established herein that the DM marker HbA1c is closely related with the CHD risk factors, so the diabetic patients should be care on CHD.