Diagnostic criteria to identify small lung adenocarcinomas that relapse after resection have yet to be established. For this purpose, we developed a mathematical logistic model in the present study. We collected data for patients with lung adenocarcinoma of 2 cm or less in size: the original cohort comprised 28 men and 25 women and the validation cohort comprised 11 men. By entering five clinicopathological factors (vascular invasion, lymphatic permeation, histological subtype, papillary carcinoma component, and smoking status) into the logistic model, we calculated a predictive function for relapse after surgery. The obtained predictive function accurately classified the patients into a recurrence or non-recurrence group: the overall accuracy of the predictive model for recurrence established from the male patients in the original cohort was 86%. Our predictive model is, however, currently limited to male patients only, because the original cohort included only one female patient with relapse. By applying the logistic model to the validation cohort, six patients were classified into a recurrence group and the other five into a non-recurrence group: four of the six patients in a recurrence group had relapsed, while all five patients in the non-recurrence group were well during their follow-up periods. Although the predictive ability of the logistic model did not reach a statistical significance (P=0.0606), nine of the 11 (82%) patients in the validation cohort were correctly classified. Consequently, using a logistic predictive model consisting of the five clinicopathological factors might enable us to predict the recurrence of resected small-sized lung adenocarcinomas in male patients.