BACKGROUND: Different quitting trajectories may reflect the stages of behavior change among smokers. The nature of quitting patterns could help the public health sector to design appropriate cessation plans. This study compared demographic, health, and behavioral characteristics and the effects of long-term abstinence between early and delayed abstainers. METHODS: We retrospectively sampled 142 abstainers from smoking cessation clinic participants for a study conducted from January 1 to December 31, 2017. Baseline information was obtained at the first clinic visit, and phone interviews were conducted 2 weeks, 3 months, and 6 months later. The 7-day point prevalence abstinence was employed for measuring their quitting status. We defined early abstainers as those who attained abstinence by second week and delayed abstainers as those who had done so by the third month. We compared their characteristics and 6-month quit rates and examined potential predictors between the two quitting patterns. RESULTS: One hundred forty-two participants were included with 87 (61.3%) early abstainers and 55 (38.7%) delayed abstainers. Early abstainers were older with more comorbidities, presenting longer smoking duration, higher exhaled carbon monoxide (CO) concentration and Fagerstrom Test of Cigarette Dependence (FTCD) scores. The 6-month abstinence rate was high for both quitting patterns with no significant difference (83.9% versus 81.8%, p = 0.7462). Higher FTCD scores and exhaled CO concentration were potential predictors for early abstainers with adjusted odds ratio 1.16 (95% confidence interval [CI], 1.01-1.33) and 1.04 (95% CI, 1.00-1.08) respectively. CONCLUSIONS: Our results associated early abstainers with older age, more comorbidities and higher nicotine dependence. Both groups achieved good long-term abstinence maintenance. Although early abstainers may achieve earlier reduction of health risks, smokers could still benefit from long-term abstinence if they can manage to quit smoking even at later phases of cessation courses.