COVID-19 has significantly disrupted global life, leading to high infection rates, considerable mortality, and economic hardships. Mitigation efforts depended on vaccination, preventive behaviors, and lifestyle changes, emphasizing the importance of understanding what influences public compliance. This study investigated COVID-19 preventive behaviors within the Thai Muslim community in Nakhon Si Thammarat Province, southern Thailand, and developed a protection model based on the World Health Organization's framework. Using a multi-stage cluster sampling method, 233 participants from eight villages completed a questionnaire adapted from CDC guidelines. Data were analyzed using Partial Least Squares Structural Equation Modeling (Smart PLS-SEM). Results demonstrated strong reliability (CR > 0.80) despite moderate AVE values (0.314–0.491); HTMT supported discriminant validity. Key predictors of compliance included community nutrition (β = 0.435, p < 0.001) and public trust (β = 0.332, p = 0.001). Additionally, the first service had a significant influence on social support (β = 0.638, p < 0.001), which subsequently affected nutritional status (β = 0.401, p < 0.001). The model explained a moderate amount of variance (R² = 0.232–0.494), showed acceptable fit (SRMR = 0.089), and demonstrated predictive validity (Q² > 0). Overall, findings emphasize the vital roles of nutrition and trust, mediated through services and social support, in encouraging compliance with health protocols.