Predicting long-term outcomes in individuals with childhood wheezing is of major clinical relevance. To identify and characterize childhood wheezing phenotypes related to asthma persistence in adolescence with a multidimensional statistical model, independent of predefined hypotheses. This prospective cohort study included 308 children, ages < 7 years, with recurrent wheezing. We systematically evaluated asthma prevalence in children at 3, 8, and 13 years of follow-ups. Risk factors associated with asthma persistence in adolescence were analyzed with multivariable logistic regression. Early childhood wheezing phenotypes were identified with k-means cluster analysis of variables selected with the logistic regression analysis, which were based on questionnaires and skin-prick tests. These phenotypes were compared for predicting asthma prevalence, use of control treatments, and lung function in childhood and adolescence. Asthma prevalence was 58.3% (n = 249) and 53.5% (n = 170) at the 8- and 13-year follow-ups, respectively. Preschool-age diagnoses of atopy (odds ratio 11.8 [95% confidence interval, 4.0-34.6]) and rhinitis (odds ratio 10.4 [95% confidence interval, 3.7-29.1]) were independent risk factors for asthma persistence in adolescence. We identified three early childhood wheezing phenotypes: transient, persistent atopic, and persistent nonatopic. The latter two were characterized by rhinitis during preschool age. These phenotypes could predict the following outcomes: asthma symptom persistence, use of control treatments, and lung function during childhood and adolescence (p < 0.03). Asthma persistence through adolescence reflected different wheezing phenotypes based on preschool-age comorbidities, particularly rhinitis, with or without atopy. Our results supported that wheezing phenotypes, identified at early ages from simple measurements, could predict asthma and lung function outcomes.