In the past few decades， water bodies around the world have continued to suffer from systemic pollution and severe water quality deterioration. Total phosphorus is one of the important indicators for water quality evaluation and an important factor affecting water eutrophication and cyanobacteria bloom outbreaks. This paper discusses the relationship between total phosphorus and other optical water quality parameters， the remote sensing inversion of total phosphorus concentration in different water body types， the remote sensing algorithm of total phosphorus concentration， and the remote sensing inversion of total phosphorus concentration on different remote sensing platforms. Since the 1990s， there have been more than 300 documents on total phosphorus concentration inversion. In recent years， research hot spots have gradually focused on topics such as “remote sensing technology” and “machine learning”. For the study of inland lake water bodies， Landsat/TM and MODIS images are mainly used. Among them， the accuracy of models using the combination of green band， near-infrared band and mid-infrared band is generally higher. Observing the total phosphorus concentration of global lakes through satellite image data， it was found that the total phosphorus content of global lakes is generally on the rise， with the highest phosphorus content in Asian lakes， followed by South America， Africa and Europe. No significant increasing trend was found in Oceania. With the development of computer technology， machine learning algorithms have gradually become a current hot topic. Compared with traditional algorithms， models built using machine learning algorithms are more accurate. The random forest algorithm is widely used because the model it builds has higher accuracy than other machine learning algorithms. With the continuous development of research， the construction of a total phosphorus concentration model suitable for different water types is the general trend in the future， and the development of sensors with high spatial resolution and high temporal resolution is even more urgent.