Impulsivity is associated with substance use, including tobacco use. The degree to which impulsivity fluctuates over time within persons, and the degree to which such intra-individual changes can be measured reliably and validly in ambulatory assessments is not known, however. The current study evaluated two novel ambulatory measures of impulsive choice and impulsive action. Impulsive choice was measured with an eight-item delay discounting task designed to estimate the subjective value of delayed monetary rewards. Impulsive action was measured with a two-minute performance test to assess behavioral disinhibition (the inability to inhibit a motor response when signaled that such a response will not be rewarded). Valid data on impulsive choice were collected at 70% of scheduled reports and valid data on impulsive action were collected on 55% of scheduled reports, on average. Impulsive choice and action data were not normally distributed, but models of relations of these measures with within- and between-person covariates were robust across distributional assumptions. Intra-class correlations were substantial for both impulsive choice and action measures. Between persons, random intercepts in impulsive choice and action were significantly related to laboratory levels of their respective facets of impulsivity, but not self-reported or other facets of impulsivity. Validity of the ambulatory measures is supported by associations between abstinence from smoking and increased impulsivity, but challenged by an association between strong temptations to smoke and reduced impulsive choice. Results suggest that meaningful variance in impulsive choice and action can be captured using ambulatory methods, but that additional measure refinement is needed.