Healthy and strong body function was essential in the job of prison management, and smoking had been proved to induce negative impacts. Based on the health belief model, the current research explored the association among health belief on smoke -cessation, self-efficacy on ceasing smoke, and possibility of ceasing smoke among the prison officials. Samples were 182 officials recruited by snowball sampling method, 72.53% were male. Test results indicated the barrier to ceasing smoke had a lowest average 2.95(±1.04), followed by benefit of ceasing smoke (3.49±0.54), perceived susceptibility (3.51±0.75), perceived severity (3.59± 0.78), self-efficacy (3.65±0.77), and possibility of ceasing smoke (3.67 ± 0.99). As to the variance analysis, history of smoking and residence were two major variables that made the constructs different, whereas age and marriage contributed the least. Noteworthy was that the education is the major contributor to the difference of smoke-ceasing possibility among respondents. Results from regression analyses indicated that perceived benefits (β=.52) and perceived barriers (β=.23) were the two major predictors of smoke-ceasing, and the other independent variables (IV) were not significant in this model (adj. R2=.279). The variance explained increased to 48.3% (adj. R2=.483) in the model with additional variable of self-efficacy. Perceived benefits and perceived barriers were replaced by the self-efficacy, and made all original IVs not significant. This indicated the mediating effects of self-efficacy between predictors and smoke-ceasing. The current research suggested the judicial correction institute shall logically staff required officials for the jobs of prison management, and recommend several smoke-ceasing measures.