This research has three critical research objectives. The first objective is to study the context and risk of diabetes from the food consumption behavior of Thai people in the northern region. The second objective is to develop a model for predicting diabetes risk from the food consumption behavior of Thai people in the northern region. The last objective is to evaluate the efficiency of the model for predicting diabetes risk from the dietary behavior of Thai people in the northern region. The population and sample in this research were 370 people in Chiang Muan District, Phayao Province, aged between 20 and 99 years. The research tools include an IOC-approved questionnaire and data mining analysis using machine learning and artificial intelligence tools known as CRISP-DM. Model performance was evaluated using data partitioning and testing known as split-validation methods and confusion matrix performance. The research results found that the context of rural people in northern Thailand toward food consumption behavior was normal. Overall, the risk of people developing diabetes was at an acceptable level. The developed model has the highest level of accuracy with an accuracy value of 81.80%. It is given in Tables 11 and 12 that people have good food consumption habits. Moreover, the model for predicting the risk of diabetes in rural communities was found to have a high level of accuracy by comparing various techniques, as given in Table 4. This research, therefore, concludes that all objectives were achieved. The results and expectations from the research are to know the food consumption context of rural people in the northern region of Thailand, which can be used to design and plan strategies to prevent the increase in people at risk of diabetes in Thailand.