Research Inclusion (RI) seeks to improve the inclusion and participation of marginalised groups with significant health care needs in research. However, research programmes are known to under-recruit individuals with multiple health problems, those facing health inequalities, and those living in precarious social situations where stigma and discrimination. 1 Although RI strategies are likely to vary by health conditions, research designs, the method of data collection, venues for data collection, and the nature and complexity of interventions, there is little practical information for researchers on how do to this. In this paper, we synthesise insights and lessons learnt from a range of research programmes in England to identify effective approaches to RI. These include studies of adolescents living with adverse childhood experiences in complex intersectional positions of vulnerability; studies of ethnicity and multimorbidity including psychosis; and research to improve public and patient involvement with ultra marginalised adolescents. The research projects sought to ensure representation of, participation by, people with lived experience, and from the most marginalised groups facing multiple forms of vulnerability. We conclude that inclusion strategies must continually evolve from the inception of research and during the delivery of the research; address power imbalances and strengthen trust through co-design and participatory methods; build community partnerships and networks of trusted organisations; adapt research infrastructures that may act as barriers to participation; ensure culturally responsive designs in order to tackle epistemic injustices when certain voices or types of knowledge are ignored. 2 Research studies should explicitly define which marginalised groups they aim to recruit, secure the necessary resources to flexibly involve people living in unexpectedly complex circumstances, revise procedures for participation and retention, transparently report successes and failures, and generate recommendations for future studies. We place our learning in the context of published literature and propose a research inclusion checklist.