Lung cancer screening may provide a "teachable moment" for the smoking cessation and relapse prevention. However, the impact of lung cancer screening on smoking initiation in non-smokers has not been reported. A baseline smoking behavior survey was conducted in 2000 participants who were screened by low-dose computed tomography (LDCT) from 2014 to 2018. All participants were re-surveyed on their smoking behavior in 2019. Of these, 312 participants were excluded, leaving 1688 participants in the final analysis. The smoking initiation rate in baseline non-smokers, the relapse rate in baseline former smokers, and the abstinence rate in baseline current smokers were calculated, respectively. The associations between screening results, demographic characteristics, and smoking behavior change were analyzed using multivariable logistic regression. From 2014 to 2019, smoking prevalence significantly decreased from 52.6% to 49.1%. The prevalence of smoking initiation, relapse, and abstinence in baseline non-smokers, former, and current smokers was 16.8%, 22.9%, and 23.7%, respectively. The risk of smoking initiation in baseline non-smokers was significantly higher in those with negative screening result (adjusted OR=2.97, 95% CI: 1.27-6.94). Compared to smokers who only received baseline screening, the chance of smoking abstinence in baseline current smokers was reduced by over 80% in those who attended 5 rounds of screening (adjusted OR=0.15, 95% CI:0.08-0.27). No significant associations were found between smoking relapse and prior screening frequency, with at least one positive screening result. Age, gender, occupational exposure, income, and smoking pack years were also associated with smoking behavior changes. The overall decreased smoking prevalence indicated an overwhelming effect of "teachable moment" on "license to smoke." A tailored smoking cessation strategy should be integrated into lung cancer screening.