Synthetic Aperture Radar (SAR) Tomography (TomoSAR) is an active microwave-based three-dimensional (3D) imaging technique that can provide large-scale, penetrative, high-resolution, all-time, and all-weather observations, making it essential for environmental monitoring such as forest above-ground biomass estimation, 3D urban reconstruction, and glacier internal structure imaging. This review systematically summarizes the four development stages of TomoSAR over the past 30 years from three aspects of systems, data processing, and applications. An analysis of 67 TomoSAR systems or projects from 23 institutions or countries gives the key characteristics and advancements of TomoSAR systems toward multi-platform, multi-configuration, multi-frequency, multi-polarization, and high-resolution. Then, three key aspects of tomographic processing—3D imaging algorithm, calibration of phase errors, and impact of temporal decorrelation – are detailed and analyzed. For the TomoSAR applications of forest, urban, and glacier scenarios, we present their scattering mechanisms using real data and explore their application potentials. Finally, how deep learning and multi-source data fusion enhance TomoSAR imaging and its applications is discussed. This review can be helpful to have a systematic understanding of TomoSAR, promoting system optimization, algorithm innovation, and application expansion.