Disaster events, such as floods, wildfires, and earthquakes, increasingly cause damage to livelihoods, the economy, and the environment. Preparing for disasters is noted as one of the most effective ways to adapt and increase resilience to these events, but research has shown that many people in the United States have not adopted recommended household preparedness actions. Moreso, there is currently no geospatial data set or tool for mapping geographic variation in disaster preparedness behavior, despite the availability of appropriate survey data. Using the Federal Emergency Management Agency National Household Survey from 2017 to 2020, we develop a multilevel regression and poststratification model that provides estimates at the state, county, and zip-code tabulation area scales of several preparedness actions and a general disaster preparedness index. Results show regional and state-level variation among preparedness levels, with the Southeast and Utah being generally more prepared than other regions of the United States. Additionally, we introduce an online interactive mapping tool for these results that practitioners, academics, and the public can use to identify preparedness levels in their area of interest. The outcomes of this study can be used to inform future work in hazard risk assessment and to further develop comparisons between risk perceptions and hazard preparedness. Finally, findings from this study contribute to the suite of geospatial models and methods used to assess the human dimensions of hazard risk and resilience.