Background: Naturalistic developmental behavioural interventions (NDBI) may improve social communication in toddlers/pre-school aged children with autism spectrum disorder (ASD). Here, we study efficacy of the low-intensity, complex NDBI ‘Frankfurt Early Intervention Program for ASD’ (A-FFIP) over 1 year by a confirmatory phase-III, prospective, randomised, controlled, parallel-group study with two treatment arms over four centres. Methods: Main inclusion criteria: ASD (DSM-5), age 24–66 months, developmental quotient >30. Intervention: Manualised A-FFIP intervention. Control intervention: Early intervention as usual (EIAU). Primary outcome: Change in core ASD symptoms from baseline (T2) to immediate intervention endpoint at 12 months (T6) based on the blindly rated Brief Observation for Communication Change (BOSCC) total score. Statistical analysis: Mixed model for repeated measures with covariates baseline BOSCC-total, chronological age and centre. Results: Between July 2018 and October 2021, N = 134 children with ASD were randomly allocated to intervention (A-FFIP: n = 68, EIAU: n = 66). Groups did not differ at baseline, with a mean age of 49 (SD 10) months, a mean developmental age of 23.3 (SD 13.6) months and 26 (19.4%) females. The SARS-CoV-2 pandemic interfered severely with trial procedures. Intention-to-treat analysis in the primary analysis set, with at least one postbaseline BOSCC measure (A-FFIP n = 64, EIAU n = 60), did not find differences in the primary outcome by group (adjusted ES −0.06, 95% CI to −0.24 to 0.11). SARS-CoV2-related lockdown led to less improvement across groups. Secondary outcomes showed stronger improvements in parent-rated repetitive behaviour as well as parent- and teacher-rated executive functions for A-FFIP versus EIAU. Adverse events were comparable between groups. Conclusions: The manualised NDBI program A-FFIP, which allows individually targeting six core basic abilities and five developmental domains related to longitudinal development in ASD, did not improve social communication, cognitive or behavioural outcomes beyond EIAU after 1 year, but may improve repetitive behaviour and executive function.