The TP concentration is an important index of water quality and an important influencing factor of eutrophication and algae blooms. Remote sensing technology has advantages of wide scope and high time limited efficacy. Monitoring the concentration of TP by satellite remote sensing is important for the study of water quality and eutrophication. In situ datasets collected during the three times of experiments in Taihu Lake between 2013 and 2014 were used to develop the TP inversion model based on GOCI data. The GOCI data in spring, summer, autumn and winter in 2014 were selected to analyze the time and space changes of TP concentration in Taihu Lake. The results showed that the TP algorithm was built up based on the variables, which was to use the eight band combination of GOCI data as variable, and build model using Multi factor linear regression method. The algorithm achieved more accurate TP estimation with R² = 0.898, MAPE = 14.296%, RMSE = 0.026 mg · L⁻¹. Meantime, a analysis on the precision of the model by using the measured sample points and the synchronous satellite images with MAPE = 33.642%, 22.551%, RMSE = 0.076 mg · L⁻¹, 0.028 mg · L⁻¹ on August 5, 2014 and October 24, 2014. Through the analysis of the 30 images on the four days of the four seasons, it showed that the absolute concentration of total phosphorus was different in different seasons. But temporal and spatial distribution of total phosphorus concentration was similar in the morning and afternoon. In spatial distribution, the TP concentration in Meiliang Bay, Zhushan Bay, Gonghu Bay, Xiaomei Port and Changdou Port in the southwest coast was at a continuously high position. The TP concentration change in different regions was influenced by wind direction, wind speed and other factors. The TP concentration highest in the morning, and then gradually decreased, this phenomenon reflected that the TP concentration was affected by temperature and light.