The associations between airway eosinophilia, measured in sputum or peripheral blood, and acute exacerbations of chronic obstructive pulmonary disease (AECOPD) are inconsistent. We therefore aimed to determine the association between eosinophilia in bronchoalveolar lavage (BAL) fluid and AECOPD in a clinical cohort. We analyzed differential cell counts from baseline BAL fluid in participants in the DISARM clinical trial (Clinicaltrials.gov #NCT02833480) and classified participants by the presence or absence of BAL eosinophilia (>1% of total leukocytes). We determined the association between BAL eosinophilia and AECOPD over 1 year of follow-up using negative binomial regression and Cox proportional hazards test. N = 63 participants were randomized, and N = 57 had BAL differential cell counts available. Participants with BAL eosinophilia (N = 21) had a significantly increased rate of acute exacerbations (unadjusted incidence rate ratio (IRR) 2.0, p = 0.048; adjusted IRR 2.24, p = 0.04) and a trend toward greater probability of acute exacerbation (unadjusted hazard ratio (HR) 1.74, p = 0.13; adjusted HR 2.3, p = 0.1) in the year of follow-up compared to participants without BAL eosinophilia (N = 36). These associations were not observed for BAL neutrophilia (N = 41 participants), BAL lymphocytosis (N = 27 participants) or peripheral blood eosinophilia at various threshold definitions (2%, N = 37; 3%, N = 27; 4%, N = 16). BAL may therefore be a sensitive marker of eosinophilic inflammation in the distal lung and may be of benefit for risk stratification or biomarker-guided therapy in COPD.