The Global Aerosol Optical Depth Product (LGHAP V2 AOD) provides daily aerosol optical depth data from 2000 to 2023 at a 1-kilometer resolution, covering terrestrial areas worldwide. Thanks to the efficient integration of multi-modal data and technological optimization, this dataset overcomes the limitations of traditional satellite data caused by cloud cover, demonstrating excellent spatiotemporal continuity and consistency. It provides key data support for air pollution monitoring, analysis of haze variation trends, and research on aerosol radiative effects, while effectively aiding ecological health assessments and climate policy formulation.The global aerosol optical depth (AOD) product relies on multi-source, multi-scale aerosol optical depth retrieval data (MODIS, MAIAC, NPP, Envisat, etc.), ground station monitoring data, as well as multi-modal geospatial big data such as meteorological and pollutant concentration numerical model simulations, and uses an advanced integrated big data analysis framework to achieve seamless global AOD generation. This method centers on tensor flow technology, combined with multi-source data fusion and missing value reconstruction strategies, and systematically handles large-scale data gaps in satellite observations through spatiotemporal tensor decomposition and principal mode reconstruction techniques. In addition, a coupled attention mechanism and background field iterative optimization algorithm are introduced to dynamically adjust data weights and background information, thereby accurately filling gaps caused by cloud cover or sensor absence. For complex terrain and meteorological conditions, the method also incorporates adaptive calibration and multi-scale spatial interpolation techniques, effectively improving reconstruction accuracy and coverage, ensuring data integrity and consistency on a global scale.