Obesity is excess fat accumulation due to an imbalance between energy intake and energy expenditure for a long time. Obesity can cause various non-communicable diseases, including heart disease, stroke, diabetes, high blood pressure, gallstones, respiratory problems, cancer, osteoarthritis, including infertility. The Regional Government through Posbindu Non-Communicable Diseases (PTM) in every UPTD Puskesmas carries out socialization as an early detection of PTM risk. This activity is carried out every two or three months, by conducting routine health checks for employees. Measurement of obesity diagnosis is usually by using the Body Mass Index (BMI). BMI is a measurement method by calculating the height in meters and weight in kilograms. However, BMI also has drawbacks, namely it cannot distinguish between muscle mass and fat mass. Apart from BMI, other factors also affect the diagnosis of obesity, including gender, age, abdominal circumference, risky behaviors such as lack of physical activity, eating patterns of excess sugar, excess salt, excess fat, not eating enough fruits and vegetables. With so many determinants of the diagnosis of obesity, the Naïve Bayes Algorithm is used to predict the diagnosis of obesity more effectively and accurately. The application built is in the form of a prototype that utilizes the PHP programming language. This study used data from the health examination of participants in the socialization of PTM risk early detection with a total of 60 participants. There are 10 variables used, namely Gender, Age, Height, Weight, Abdominal Circumference, Lack of Physical Activity, Eating Patterns of Excess Sugar, Excess Salt, Excess Fat and Undereating and Fruit while there are 2 classes, namely obesity and normal diagnoses. Based on the results of calculating accuracy with the Confusion Matrix, an accuracy value of 86.6% is obtained.