According to Prochaska's transtheorical model, the stage of action is a milestone in the long-term smoking cessation process. Describing smokers' variables preceding this stage is essential. Methods: A retrospective case-study design using data from the FAMUS primary care register. The objective was to develop a predictive model for the smoker's action of quitting smoking. The quit-attempt rate (QAR) is defined as the proportion of patients who attempted to quit at least once during the follow-up period. Tree-based and logistic regression models were applied to 7,567 smokers with at least two recorded visits. Results: The QAR is 25% and is higher for men (26.5%) than for women (24.5%). For men, the groups generated by the trees are light (≤3 cigarettes), moderate (4-15), or heavy smokers (>15). For women, the pattern observed involves four groups: very light (1 cigarette), moderate-low (2-10), moderate-high (11-20), and heavy smokers (>20). In both cases, the QAR decreases from light to heavy smokers, and the number of cigarettes, nicotine replacement therapy, and age remain relevant factors predicting the QAR. Conclusions: Medical professionals should consider smoking an addiction and take specific approaches to smoking cessation based on age, gender, and cigarettes consumed.