Abstract Two ecological models that describe phosphorus-food web dynamics with and without macrophytes in Lake Chao ecosystems are developed in this paper. Ecosystem health indicators used in the two models include exergy ( Ex), structural exergy ( Ex st), the ratio of zooplankton to phytoplankton biomass ( R BZBA), and transparency in Secchi Disc depth ( SD). The calculated values of important state variables and process rates from model 1 are compared with the measured data. Model 2 uses the same parameters and rate coefficients as model 1 for phytoplankton, zooplankton, fish, detritus, sediment, and soluble inorganic phosphorus submodels. The macrophyte submodel in model 2 is given realistic parameters taken from the literature. The lake ecosystem health indicators are calculated and compared by means of models 1 and 2. The changes of ecosystem health indicators and lake biological structure following macrophyte restoration are predicted using model 2. The results from model 1 show that there are good agreements between observed and simulated values of important variables and process rates. The results from model 2 reveal that macrophyte restoration in the Lake Chao ecosystem can decrease phytoplankton biomass, and increase Ex, Ex st, SD, R BZBA, and fish biomass. With the increase of initial macrophyte biomass, phytoplankton biomass is declined, and Ex, Ex st, SD, R BZBA, and fish biomass are increased. These findings mean that macrophyte restoration can regulate lake biological structure and improve lake ecosystem health.