Background: Psychotherapy for borderline personality disorder is often extensive and resource-intensive. Mentalisation-based therapy is a psychodynamically oriented treatment option for borderline personality disorder, which includes a case formulation, psychoeducation, and group and individual therapy. The evidence on short-term compared with long-term mentalisation-based therapy is currently unknown. Methods/design: The Short-Term MBT Project (MBT-RCT) is a single-centre, parallel-group, investigator-initiated, randomised clinical superiority trial in which short-term (20 weeks) will be compared with long-term (14 months) mentalisation-based therapy for outpatients with subthreshold or diagnosed borderline personality disorder. Outcome assessors, data managers, the data safety and monitoring committee, statisticians, and decision-makers will be blinded to treatment allocation. Participants will be assessed before randomisation and at 8, 16, and 24 months after randomisation. The primary outcome will be the severity of borderline symptomatology assessed with the Zanarini Rating Scale for Borderline Personality Disorder. Secondary outcomes will be functional impairment (Work and Social Adjustment Scale), quality of life (Short-Form Health Survey 36—mental component), global functioning (Global Assessment of Functioning), and proportion of participants with severe self-harm. In this paper, we present a detailed statistical analysis plan including a comprehensive explanation of the planned statistical analyses, methods to handle missing data, and assessments of the underlying statistical assumptions. Final statistical analyses will be conducted independently by two statisticians following the present plan. Discussion: We have developed this statistical analysis plan before unblinding of the trial results in line with the Declaration of Helsinki and the International Conference on Harmonization of Good Clinical Practice Guidelines, which should increase the validity of the MBT-RCT trial by mitigation of analysis bias. Trial registration: ClinicalTrials.gov NCT03677037. Registered on 19 September 2018