Mothers’ knowledge about childhood asthma influences management practices and disease control, but validating knowledge/practice questionnaires is difficult due to the lack of a gold standard. We hypothesized that Latent Class Analysis (LCA) could help identify underlying mother profiles with similar knowledge/practices. A total of 438 mothers of asthmatic children answered a knowledge/practice questionnaire. Using answers to the knowledge/practice questionnaire as manifest variables, LCA identified two classes: Class 1, “poor knowledge” (33%); Class 2, “good knowledge” (67%). Classification accuracy was 0.96. Mothers in Class 2 were more likely to be aware of asthma-worsening factors and indicators of attacks. Mothers in Class 1 were more likely to prevent exposure to tobacco smoke (91.1% vs. 78.8%, p = 0.005). For attacks, mothers in Class 2 were more likely to go to the emergency department and follow the asthma action plan. Mothers in Class 2 more frequently had a high education level (79.5% vs. 65.2%, p = 0.004). Children in Class 2 more frequently had fully controlled asthma (36.7% vs. 25.9%, p = 0.015) and hospitalizations for attacks in the previous 12 months (24.2% vs. 10.7%, p = 0.003). LCA can help discover underlying mother profiles and plan targeted educational interventions.