Measuring classroom behavior among young children is important to guide assessment and intervention decisions, yet there is limited literature on appropriate direct observation tools for this purpose. This article describes the psychometric properties of the Behavior Assessment System for Children, Student Observation System (BASC-3 SOS) with 135 children ages 20 to 67 months (M = 35 months, 64% Latinx, 78% with an established developmental disability) and their teachers (N = 36) as part of a larger randomized control trial of a teacher training intervention. Inter-rater reliability on individual BASC-3 SOS behaviors ranged from poor to good. Correlations between BASC-3 SOS scores across time indicated low to moderate developmental test–retest reliability. Significant correlations between BASC-3 SOS scores and teacher ratings provided evidence for convergent, divergent, and predictive validity. Differences between BASC-3 SOS scores for children with versus without disabilities supported the tool’s discriminant validity. There were no significant pre- to post-treatment changes in BASC-3 SOS scores. Overall, results provide mixed evidence for the psychometric properties of the BASC-3 SOS when used with young, diverse children with and without disabilities. Implications for clinical and research purposes are discussed.