AIM:To compare language, speech, and voice of children born preterm and at term, and determine relevant predictors of outcome. METHOD:Three hundred infants (150 males, 150 females; 149 born at <30wks' gestation, 151 term-born) were prospectively recruited at birth from the Royal Women's Hospital. We administered the Preschool Language Scales, Fifth Edition, Diagnostic Evaluation of Articulation and Phonology, Grade Roughness Breathiness Asthenia Strain Scale, and Pediatric Voice Handicap Index at 3 years, and compared groups. We examined hypothesized predictors in children born preterm: gestational age at birth, birthweight, sex, chronic lung disease, high social risk, multilingualism, neurodevelopmental diagnosis, and oromotor feeding. RESULTS:Children born preterm had poorer language than children born at term (coefficient -5.43). Speech and voice were similar between groups (coefficients -0.70 to 1.63). Chronic lung disease predicted voice (coefficient 6.05); male sex (coefficients 4.54-6.18), high social risk (coefficient -6.02 to -9.30), and neurodevelopmental diagnosis (coefficients -16.42 to -20.61) predicted language. INTERPRETATION:Children born before 30 weeks' gestation had poorer language than children born at term. Children born preterm with neurodevelopmental disabilities or high social risk experience poorer language outcomes, and would benefit from enrichment of their language environment. WHAT THIS PAPER ADDS:Speech and voice outcomes were similar between children born preterm and at term. Male sex, high social risk, and neurodevelopmental diagnosis predicted language outcomes.