Rationale: Children with preschool wheezing represent a very heterogeneous population with wide variability regarding their clinical, inflammatory, obstructive, and/or remodeling patterns. We hypothesized that assessing bronchial remodeling would help clinicians to better characterize severe preschool wheezers. Objectives: The main objective was to identify bronchial remodeling-based latent classes of severe preschool wheezers. Secondary objectives were to compare cross-sectional and longitudinal clinical and biological data between classes and to assess the safety of bronchoscopy. Methods: This double-center prospective study (NCT02806466) included severe preschool wheezers (1-5 yr old) requiring fiberoptic bronchoscopy. Bronchial remodeling parameters (i.e., epithelial integrity, reticular basement membrane [RBM] thickness, mucus gland, fibrosis and bronchial smooth muscle [BSM] areas, the density of blood vessels, and RBM-BSM distance) were assessed and evaluated by latent class analysis. An independent cohort of severe preschool wheezers (NCT04558671) was used to validate our results. Measurements and Main Results: Fiberoptic bronchoscopy procedures were well tolerated. A two-class model was identified: Class BR1 was characterized by increased RBM thickness, normalized BSM area, the density of blood vessels, decreased mucus gland area, fibrosis, and RBM-BSM distance compared with Class BR2. No significant differences were found between classes in the year before fiberoptic bronchoscopy. By contrast, Class BR1 was associated with a shorter time to first exacerbation and an increased risk of both frequent (3 or more) and severe exacerbations during the year after bronchoscopy in the two cohorts. Conclusions: Assessing bronchial remodeling identified severe preschool wheezers at risk of frequent and severe subsequent exacerbations with a favorable benefit to risk ratio.