Functional magnetic resonance imaging (fMRI), a non-invasive technique based on the blood oxygen-level dependent (BOLD) hemodynamic principle, has emerged as a widely used tool to map brain activity. With its balanced spatial and temporal resolution, high repeatability, and whole-brain imaging capabilities, fMRI has greatly contributed to our understanding of the brain mechanism underlying health and disease. It has become an indispensable technique in large-scale multi-center projects aimed at investigating various perspectives. The explosive growth of fMRI studies has significantly increased our knowledge of the brain, leading to the generation of large brain datasets and the accumulation of published findings, thereby moving the field into the era of “big data.” This shift presents unprecedented opportunities for discovery in brain science but also introduces various challenges, such as data acquisition, management, and analysis. In this chapter, we present a comprehensive review of fMRI studies from a big data perspective. We highlight recent advances in state-of-the-art data acquisition with critical features, including high spatial and/or temporal precision, large sample sizes, long-term recording of brain activity, multidimensional biological variables (e.g., imaging, genetic, demographic, cognitive, and clinical), and vast quantities of existing findings. We then review methodological advances in fMRI big data analysis, such as multi-center harmonization, graph theoretical analysis, dynamic network analysis, multivariate pattern analysis, and machine learning. Moreover, we discuss recent findings in the application of fMRI big data to the exploration of functional network architecture that underlies cognition and behavior, normal development and aging, and neuropsychiatric disorders. Finally, we discuss emerging challenges and issues that need to be resolved urgently in this field; we hope to provide a roadmap for future investigations and inspire new breakthroughs in the field of fMRI research.