Frequent outbreaks of cyanobacterial blooms in Lake Taihu are undoubtedly a great threat to the economic development of its neighboring areas and the safety of drinking water of its residents. This paper takes Taihu Lake as the study area and analyzes its geographic location information and development status. Then, based on the remote sensing data from MODIS and Landsat 8 satellites, the normalized vegetation index is improved to identify the blooms, and the dynamic detection method of cyanobacterial blooms is constructed by combining with the remote sensing inversion of water temperature. At the same time, the spectral performance of each band is integrated to excavate the characteristic information of cyanobacterial bloom, and the algorithm in this paper is used to process the satellite remote sensing data of cyanobacterial bloom in Lake Taihu to analyze its spatial and temporal distribution characteristics, which is used as the basis of the dynamic warning model for early warning. Then the LightGBM method is introduced to realize the all-weather spatial and temporal continuous monitoring of cyanobacterial blooms in Lake Tai. Analyzing the monitoring data of this paper’s model on the intraday change process of cyanobacterial bloom in Lake Taihu, it is found that the trend of intraday change in the area of cyanobacterial bloom in Lake Taihu in different seasons is relatively consistent, with the highest area of the bloom in autumn, accounting for 21% of the area of Lake Taihu’s water body. The study pointed out that after entering the fall, extra attention should be paid to the monitoring, prevention and control of cyanobacterial bloom in Lake Taihu.