Context: The literature on self-management interventions (SMIs) is growing exponentially, but it is characterized by heterogeneous reporting that limits comparability across studies and interventions. Building an SMI taxonomy is the first step towards creating a common language for stakeholders to drive research in this area and promote patient self-management and empowerment. Objective: To develop and validate the content of a comprehensive taxonomy of SMIs for long-term conditions that will help identify key characteristics and facilitate design, reporting and comparisons of SMIs. Methods: We employed a mixed-methods approach incorporating a literature review, an iterative consultation process and mapping of key domains, concepts and elements to develop an initial SMI taxonomy that was subsequently reviewed in a two-round online Delphi survey with a purposive sample of international experts. Results: The final SMI taxonomy has 132 components classified into four domains: intervention characteristics, expected patient/caregiver self-management behaviours, outcomes for measuring SMIs and target population characteristics. The two-round Delphi exercise involving 27 international experts demonstrated overall high agreement with the proposed items, with a mean score (on a scale of 1-9) per component of 8.0 (range 6.1-8.8) in round 1 and 8.1 (range 7.0-8.9) in round 2. Conclusions: The SMI taxonomy contributes to building a common framework for the patient self-management field and can help implement and improve patient empowerment and facilitate comparative effectiveness research of SMIs. Patient or public contribution. Patients’ representatives contributed as experts in the Delphi process and as partners of the consortium.