Abstract Cardiac rehabilitation (CR) programs are effective at improving cardiovascular outcomes by helping patients make heart-healthy lifestyle changes. However, these gains often do not last since as many as 50% of CR patients stop exercising or eating a heart-healthy diet within a few months of completing a CR program. Mobile health (mHealth) has the potential to provide effective ongoing support after CR programs end, and to help patients sustain and augment health gains obtained during CR. However, to realize this promise, new approaches are needed to increase the engagement in mHealth interventions. Patients abandon these interventions for many reasons, including because they get discouraged when they don't meet their goals, frustrated when they feel the mHealth system is sending them messages at wrong times or too often, or when other demands in their lives take priority over trying to be active or eat healthily. Once abandoned mHealth interventions stop being able to provide any further support. In this project, we will address this important problem by developing novel algorithms and user interfaces (elements of an mHealth app individuals use to interact with the app) that will help mHealth interventions to effectively adapt to the changes in individuals' needs and priorities, so the intervention can remain useful—and thus have an opportunity to support patients— over the long-term. The methods will (1) help mHealth systems more effectively provide interventions by monitoring and optimizing how these interventions impact both intermediate-term behavioral outcomes like commitment to physical activity—which mediate long-term change—as well as engagement with the intervention itself; and (2) enable users to directly make changes to how an intervention behaves by examining and correcting the information that the system uses to make decisions about intervention provision and specifying the level of support that better matches their current circumstances and priorities. Once developed and optimized in a micro-randomized trial with 60 CR patients, we will evaluate the ability of these innovations to improve engagement with an mHealth intervention for physical activity and to improve physical activity itself in a 9-month randomized controlled trial with 150 CR patients. The trial will compare an intervention that contains the newly developed algorithms and interfaces with one that includes the same components for supporting behavior change but which does not include the new algorithms and interfaces for increasing engagement. If found to be successful, the innovations developed in this project will enable the development of a new generation of mHealth interventions that can effectively support behavior change and maintenance both for cardiovascular risk reduction and in other domains of health.