Just-in-Time Adaptive Interventions (JITAIs) are designed to tailor and adapt the provision of intervention strategies (e.g., the type, timing, intensity) to an individual’s changing status and contexts with the goal to deliver support at the moment and in the context a person needs it most and is most likely to be receptive. JITAIs are envisioned to promote long-term beneficial health outcomes (distal outcomes) through targeting behavior change mechanisms that can be influenced in-the-moment (proximal outcomes). Despite JITAIs’ promise, missing data in proximal outcomes in Micro-Randomized Trials (MRT) impedes building the evidence-base necessary to inform their development. An MRT is an experimental design whose typical primary aims (e.g., to assess an average main effect on a proximal outcome of interest) and exploratory aims (e.g., whether the main effect varies over time) may inform which intervention components to include in an intervention package and contexts in which they may be effective. An MRT is characterized by frequent decision points, i.e., moments of time when micro-randomization of an individual to intervention component options may or may not occur. For instance, in the Mobile Assistance for Regulating Smoking (MARS) study, a 10-day MRT promoting real-time real-world engagement with self-regulatory strategies among smokers attempting to quit, micro-randomizations did not occur when a participant was driving and delivery of ecological momentary assessments were contingent on micro-randomization occurring. Restrictions on momentary eligibility for micro-randomization, as exemplified by the MARS study, results in patterns of micro-randomization eligibility over time (e.g., some combination of eligible or ineligible for micro-randomization up to the current time). Although momentary ineligibility for micro-randomization does not produce missing data in proximal outcomes, it does impede a direct application of Sequential Regression Multivariate Imputation (SRMI), a popular regression-based approach to multiple imputation (also known as Multiple Imputation by Chained Equations or MICE). We extend SRMI to impute missing observations stratified by their patterns of micro-randomization eligibility and respecting the temporal order of MRT data collection. We illustrate the utility of our approach to building an evidence-base for a JITAI with typical primary aims and exploratory aims of an MRT and provide an application to the MARS study.