The largest type of organized convective storm, known as the mesoscale convective system (MCS), plays an important role in the global hydrological and energy cycles. In this study, we develop a new methodology to construct a global high-resolution (~10 km, hourly) MCS tracking database using the merged global geostationary infrared temperature (T_b) and the Integrated Multi-satellitE Retrievals for GPM (IMERG) V06 precipitation datasets. The new method tracks and identifies MCS based on both T_b and precipitation feature (PF) characteristics, an improvement compared to previous works that only use T_b. Independent validations against ground-based radar network observations in the U.S. and China show that the satellite-based MCS dataset agree well with ground-based observations in terms of MCS frequency, precipitation amount, diurnal cycle and PF characteristics. This is the first global MCS tracking database that covers both the tropics and midlatitudes for all seasons. Results show that MCS accounts for over 50% of annual total rainfall across the majority of the tropical belt (15°S – 15°N), and in select regions of the subtropics and midlatitudes such as the South Asia monsoon region, North and South Americas, and oceanic regions east of the midlatitude continents. The longest-lived MCSs preferentially occur over the subtropical oceans. Compared to tropical MCSs, midlatitude MCSs have much larger and faster propagating PFs, particularly over the ocean. While land MCSs have deeper cloud-tops associated with more intense convective updrafts, oceanic MCSs have much higher rainfall intensity and heavier rainfall volume ratio. These results suggest while MCSs are ubiquitous around the globe, there are fundamental differences in their dynamic and thermodynamic structures that warrant further studies to better understand processes that control their evolutions.