OBJECTIVE: To investigate the processes of change, demographic, health- and smoking-related predictors of both smoking cessation and smoking reduction in adolescents. METHODS: Data were drawn from a sample of 755 adolescent smokers who participated in a study testing the efficacy of a text messaging-based intervention for smoking cessation. Demographic, health- and smoking-related variables were assessed at baseline. Five processes of smoking cessation, derived from the Transtheoretical Model and the Social Cognitive Theory, as well as outcome measures were assessed at 6-month follow up. Univariate and multivariate regression analyses were conducted to identify baseline and process variables to predict smoking abstinence and smoking reduction. RESULTS: Male gender (OR=0.43, p<.01), lower alcohol consumption (OR=0.90, p=.05) and a lower number of cigarettes smoked per day at baseline (OR=0.87, p<.01) predicted smoking abstinence. Baseline physical activity predicted smoking reduction (OR=1.04, p=.03). None of the examined process variables significantly predicted smoking abstinence. The process variable "counter-conditioning" predicted smoking reduction (OR=1.46, p=.03). CONCLUSIONS: Baseline predictors of smoking cessation differ from predictors of smoking reduction. Dynamic or modifiable variables play an important role in predicting adolescent smoking cessation. PRACTICE IMPLICATIONS: Counter-conditioning might be an important element in adolescent smoking cessation interventions.