Background Child and adolescent psychopathology has been linked to increased sleep problems, but there has been less investigation of this relationship in younger samples with early‐onset psychopathology. This study examined three specific but commonly observed aspects of sleep behaviors in young children – (i) Sleep onset latency, (ii) Refusal to sleep alone, and (iii) Nighttime awakenings – measured during preschool, and investigated whether these sleep problems predicted anxiety and/or depression across the next 6 years until school age (ages 9–13). Methods Data were analyzed from N = 292 participants from a prospective longitudinal study of preschool‐age children (ages 3–6). At baseline, parent‐reported clinical interviews of psychiatric symptoms, as well as sleep problems were conducted using the Preschool‐Age Psychiatric Assessment (PAPA). Follow‐up clinical interviews were also conducted annually through school age using the Childhood and Adolescent Psychiatric Assessment (CAPA). Results Parent‐reported sleep onset latency and refusal to sleep alone were significant independent predictors of MDD and anxiety severity, but not ADHD severity across time, even after controlling for family income‐to‐needs ratio and maternal internalizing psychopathology. In exploratory analyses using only healthy preschoolers, parent‐reported sleep onset latency and refusal to sleep alone also predicted anxiety severity. Conclusions We demonstrate that specific, yet relatively common sleep problems predict diagnostic severity of depression and anxiety across time, but not ADHD. Increased clinical attention to and screening for sleep onset latency and refusal to sleep alone during preschool may be warranted.