Correlative light and electron microscopy (CLEM) combines the strengths of both light and electron imaging modalities and enables linking of biological spatiotemporal information from live-cell fluorescence light microscopy (fLM) to high-resolution cellular ultra-structures from cryo-electron microscopy and tomography (cryoEM/ET). This has been previously achieved by using fLM signals to localize the regions of interest under cryogenic conditions. The correlation process, however, is often tedious and time-consuming with low throughput and limited accuracy, because multiple correlation steps at different length scales are largely carried out manually. Here, we present an experimental workflow, AutoCLEM, which overcomes the existing limitations and improves the performance and throughput of CLEM methods, and associated software. The AutoCLEM system encompasses a high-speed confocal live-cell imaging module to acquire an automated fLM grid atlas that is linked to the cryoEM grid atlas, followed by cryofLM imaging after freezing. The fLM coordinates of the targeted areas are automatically converted to cryoEM/ET and refined using fluorescent fiducial beads. This AutoCLEM workflow significantly accelerates the correlation efficiency between live-cell fluorescence imaging and cryoEM/ET structural analysis, as demonstrated by visualizing human immunodeficiency virus type 1 (HIV-1) interacting with host cells.