The study objectives are to verify a path analysis of mental health among Thai elderly with diabetes mellitus. The sampling method used is stratified random sampling. The sample size comprises 2,000 Thai elderly persons over 60 years old with diabetes mellitus. The variables are divided into individual (food, exercise, BMI and drugs), psychological (illness perception) and social factors (family support). Data are analysed using path analysis. The total variables described mental health change by 64.5%. It was found that exercise, illness perception, food and drugs as variables had a direct influence on mental health with standardised regression weights of 0.732, 0.347, 0.243 and 0.050, respectively (p-value<0.01). The exercise variable had a direct effect on mental health the most. The benefits of this study include providing recommendations to guide future mental health policy and to reduce the cost of health maintenance in Thailand.