Distinct phenotypes can be identified in childhood wheezing illness. Within the context of a birth cohort study, we investigated the association between preschool lung function and phenotypes of wheeze. From parentally reported history of wheeze (interviewer-administered questionnaire, age 3 and 5 years), children were classified as never wheezers, transient early wheezers, late-onset wheezers, or persistent wheezers. Lung function (specific airway resistance [sRaw]; kPa/second) was assessed at age 3 (n = 463) and 5 years (n = 690). Persistent wheezers had markedly poorer lung function compared with other groups. In children who had wheezed by age 3, the risk of persistent wheeze increased with increased sRaw (odds ratio [OR] 5.2, 95% confidence interval [CI] 1.3-22.0; p = 0.02). In a multivariate model, increasing sRaw (OR 5.5, 95% CI 1.2-25.9; p = 0.03) and the child's sensitization (OR 2.8, 95% CI 1.3-5.8; p = 0.008) were significant independent predictors of persistent wheezing. We found no association between lung function at age 3 and late-onset wheeze in children who had not wheezed previously (OR 0.6, 95% CI 0.07-5.3; p = 0.64). In conclusion, poor lung function at age 3 predicted the subsequent persistence of symptoms in children who had wheezed within the first 3 years, but was not associated with the onset of wheeze after age 3 in children who had not wheezed previously.