BACKGROUND: Despite autism being a lifelong developmental condition affecting approximately 2% of children and young people worldwide, interventions aimed at improving core autism features are sparse. Paediatric Autism Communication Therapy (PACT) is among the first parent-mediated developmental interventions, provided in naturalistic settings, to show promising results in core feature improvement. METHODS: DAN-PACT is an investigator-initiated, independently funded, multicentre, parallel group superiority, randomised clinical trial, which aims to assess benefits and harms of PACT in 2.0-6.9-year-old children with a recent autism diagnosis, comparing PACT combined with management as usual to management as usual alone. Two hundred eighty autistic children from all regions of Denmark will be included. Primary outcome assessors, data managers, statisticians, and conclusion drawers will be blinded. The primary outcome is magnitude of autism features as measured by the ADOS-2 CSS. The sample size calculation is based on a minimal important difference of 0.66 points, corresponding to a 2-3-point difference in ADOS raw scores. Secondary outcomes are changes in child social communication skills measured with the BOSCC, child adaptive skills measured with the VABS-3, and parents' assessment of their own and their child's quality of life. Several exploratory outcomes will be assessed, including adverse events during the trial period. Trial staff will be trained to perform both PACT and enhanced management as usual, assessing manual fidelity of PACT and measuring a broad range of benefits and harms with repeated measures. DISCUSSION: DAN-PACT aims to minimise risks of bias, anchor the trial in a naturalistic clinical setting, and extend the scope of outcome measures used in previous PACT studies, achieved by blinding raters to all observational outcomes, including a large sample of young autistic children, and using an enhanced management as usual control sample. Trial limitations are the risk of missing data and the inability to blind parent participants to their group allocation and their rating of several secondary and exploratory outcomes. In addition, there is no consensus on the magnitude of the minimal important difference on ADOS-2 CSS or BOSCC, which are therefore estimated pragmatically. TRIAL REGISTRATION: ClinicalTrials.gov NCT05673096. Registered on December 22, 2022.