Marine ciliates, as a pivotal subgroup of microzooplankton, are major consumers of primary production and important food sources for mesozooplankton, thereby channeling energy and nutrients from primary producers to higher trophic levels. Their rapid responses to environmental fluctuations, coupled with their diverse trophic modes, make them key drivers of biogeochemical cycles and sensitive bioindicators of marine ecosystem health. However, global observations of ciliate biomass remain sparse and unevenly distributed, impeding a comprehensive understanding of their spatiotemporal distribution patterns. To address this gap, we compiled a global dataset of ciliate biomass along with corresponding environmental variables and developed a stacked ensemble machine learning model that integrates multiple widely used algorithms to model the global distribution of ciliate biomass. We further projected potential changes in ciliate biomass under different climate scenarios, which can serve as reference data for comparing with and constraining predictions from ecosystem and biogeochemical models of marine biomass and associated carbon fluxes. In addition, we applied SHapley Additive exPlanations (SHAP) to quantify feature contributions and explore the environmental mechanisms governing the spatiotemporal distribution patterns. The nonlinear relationships revealed by SHAP-based partial dependence analyses further provide a basis for applying structural equation modelling (SEM) to explore potential causal pathways and to guide the parameterization of mechanistic models. Overall, our study provides the first global map of marine ciliate biomass showing their spatiotemporal characteristics, offers deep insights into the environmental drivers of their spatiotemporal variability, and advances our understanding of ciliate-mediated food web processes and their potential shifts under future ocean warming.