Mesozooplankton, defined as zooplankton ranging from 200 μm to 2cm, are a crucial link between primary producers and higher trophic level consumers, and play a vital role in the marine food web, biological carbon pump, and sustaining fishery resources. To improve our ability to predict the mesozooplankton biomass in the global ocean and explore the underlying mechanisms, we compared four machine learning algorithms (boosted regression trees, random forest, artificial neural network, and supper vector machine) to estimate the global distribution of mesozooplankton biomass. The four algorithms were built on a compilation of published mesozooplankton biomass observations (n= 130,922) with corresponding environmental predictors from contemporaneous satellite observations. The inputs of the algorithms contain spatiotemporal (sampling location, depth, date, and time) and environmental variables (temperature, chlorophyll, salinity, and oxygen concentration at surface level, and mixed layer depth). Among the four algorithms, the Random Forest (RF) achieves the best prediction performance with R² and root mean standard error (RMSE) of 0.60 and 0.39, respectively. The model outputs reveal that the Chl a concentration representing the prey availability is one of the most important factors driving the global distribution of mesozooplankton biomass. Our results show a strong seasonal cycle in mesozooplankton biomass, with higher biomass in summer and lower in winter in both northern and southern hemispheres. Also, the mesozooplankton biomass is higher at night than in the daytime in most regions due to the diel migration. The sensitivity tests by changing one predictor at one time show that mesozooplankton biomass did not change unidirectionally, suggesting the interactions among all environmental variables. We further predict the mesozooplankton biomass under the “business-as-usual” scenario and found a general decrease trend consistent with Chl a concentration. This study has enhanced our ability to predict the global distribution of mesozooplankton biomass and advance our understanding of the mechanisms underlying such global distribution.