Background: While practitioners are increasingly asked to be mindful of the evidence base of intervention programmes, evidence from rigorous trials for the effectiveness of interventions that promote oral language abilities in the early years is sparse. Aims: To evaluate the effectiveness of a language intervention programme for children identified as having poor oral language skills in preschool classes. Methods & Procedures: A randomized controlled trial was carried out in 13 UK nursery schools. In each nursery, eight children (N = 104, mean age = 3 years 11 months) with the poorest performance on standardized language measures were selected to take part. All but one child were randomly allocated to either an intervention (N = 52) or a waiting control group (N = 51). The intervention group received a 15-week oral language programme in addition to their standard nursery curriculum. The programme was delivered by trained teaching assistants and aimed to foster vocabulary knowledge, narrative and listening skills. Outcomes & Results: Initial results revealed significant differences between the intervention and control group on measures of taught vocabulary. No group differences were found on any standardized language measure; however, there were gains of moderate effect size in listening comprehension. Conclusions & Implications: The study suggests that an intervention, of moderate duration and intensity, for small groups of preschool children successfully builds vocabulary knowledge, but does not generalize to non-taught areas of language. The findings strike a note of caution about implementing language interventions of moderate duration in preschool settings. The findings also highlight the importance of including a control group in intervention studies.