Hurricanes can have devastating impacts on coastal populations in the US and across the world. To mitigate the potential impacts of hurricanes, emergency personnel need to prepare detailed evacuation plans to ensure the population has adequate time to evacuate in the event of a hurricane. The traditional approach to collecting evacuation behavior information involved conducting surveys, which are time-consuming, costly, and increasingly suffer from lowering response rates. In this study, we document a novel approach that relies on mobile location data and high-resolution parcel data to analyze evacuation decision processes such as decision to evacuate (Yes/No), and evacuation destination type (family/friends, hotel/motel, public shelter and out-of-state). The mobile location data were used to analyze the evacuation travel patterns across two major hurricanes in Florida in 2016 and 2017. The location data was augmented with high-resolution land use data sourced from Florida Department of Revenue to determine the land-use patterns to understand the home and destination characteristics of the devices. Using processed device data, we estimate evacuation participation rate and evacuation destination type (or refuge) rates by multiple variables including evacuation zone type, residential land use type, median income, and rural/urban location. Further, we employ a linear regression model system to model participation rate (per thousand) and a multinomial logit fractional split model to understand evacuation destination choice behavior. The results from the estimated models indicate that evacuation zone type, hurricane characteristics, zonal level demographic characteristics, and land use characteristics are key determinants of county-zone level participation rate and refuge rate.