Abstract The proposed R34 will develop a theory-driven, adaptive text messaging intervention (TMI) for risky drinking in postpartum women. Risky drinking is prevalent among women of reproductive age, and more than half of women who reduce their alcohol use during pregnancy return to pre-pregnancy levels by three months postpartum. Postpartum risky drinking can impair a mother’s ability to care for her baby, and is associated with increased risk for child abuse and neglect and involvement with the child welfare system, as well as poor long- term child outcomes. Most adults who engage in risky drinking do not seek treatment, and postpartum women may be especially unlikely to seek formal treatment due to stigma and fears of child protective services involvement. Text messaging is a promising strategy for reaching non-treatment-seeking risky drinkers that may be particularly suitable for postpartum women, as it is convenient and anonymous. This approach also has wide scalability potential, as it is capable of reaching large numbers of people at relatively low cost. The proposed study will apply the Multiphase Optimization Strategy (MOST) framework, an engineering-inspired approach for developing efficient and scalable mobile health interventions, to develop the first adaptive TMI specifically designed for risky drinking in postpartum women. The proposed study will refine a theoretical model to inform the TMI that is based on the Relapse Prevention model and specifies the temporal relations of stress, negative mood, and maternal self-efficacy with intention to drink and daily drinking behavior in postpartum women, and their daily fluctuations in response to triggers that are unique to the postpartum period (sleep quality and baby irritability). The proposed TMI will include behavior change interventions based on motivational interviewing and cognitive behavioral therapy that target coping skills and motivation to avoid drinking, and that are adaptively tailored to fluctuations in stress, negative mood, and self-efficacy. Specific aims are to (1) test and refine the proposed theoretical model via Ecological Momentary Assessment with 30 women for 14 days to determine daily fluctuations in stress, negative mood, and self-efficacy in response to triggers and their impact on drinking intentions; (2) develop the content and decision rules for the TMI via iterative cycles of user feedback collected via focus groups with 30 postpartum women; and (3) conduct a pilot micro-randomized trial (MRT) with 50 women to assess feasibility of design and methods and obtain preliminary proof-of-concept data to prepare for a future fully-powered optimization trial. Study products will be a set of preliminary TMI components for postpartum risky drinking, ready for optimization and evaluation in a future larger scale R01, and a set of piloted EMA and MRT methods for the population of postpartum risky drinkers.