Abstract Background The therapy of chronic musculoskeletal pain (CMSP) is complex and the treatment results are often insufficient despite numerous therapeutic options. While individual patients respond very well to specific interventions, other patients show no improvement. Personalized treatment assignment offers a promising approach to improve response rates; however, there are no validated cross-disease allocation algorithms available for the treatment of chronic pain in validated personalized pain interventions. This trial aims to test the feasibility and safety of a personalized pain psychotherapy allocation with three different treatment modules and estimate initial signals of efficacy and utility of such an approach compared to non-personalized allocation. Methods This is a randomized, controlled assessor-blinded pilot trial with a multifactorial parallel arm design. CMSP patients (n = 105) will be randomly assigned 1:1 to personalized or non-personalized treatment based on a cluster assignment of the West Haven-Yale Multidimensional Pain Inventory (MPI). In the personalized assignment condition, patients with high levels of distress receive an emotional distress-tailored intervention, patients with pain-related interference receive an exposure/extinction-tailored treatment intervention and patients who adapt relatively well to the pain receive a low-level smartphone-based activity diary intervention. In the control arm, patients receive one of the two non-matching interventions. Effect sizes will be calculated for change in core pain outcome domains (pain intensity, physical and emotional functioning, stress experience, participant ratings of improvement and satisfaction) after intervention and at follow-up. Feasibility and safety outcomes will assess rates of recruitment, retention, adherence and adverse events. Additional data on neurobiological and psychological characteristics of the patients are collected to improve treatment allocation in future studies. Conclusion Although the call for personalized treatment approaches is widely discussed, randomized controlled trials are lacking. As the personalization of treatment approaches is challenging, both allocation and intervention need to be dynamically coordinated. This study will test the feasibility and safety of a novel study design in order to provide a methodological framework for future multicentre RCTs for personalized pain psychotherapy. Trial registration German Clinical Trials Register, DRKS00022792 (https://www.drks.de). Prospectively registered on 04/06/2021.