Mathematical modeling is an important tool in understanding the structure and functioning of aquatic environments, allowing to estimate and simulate the magnitude of the impacts that activities such as fish farming have on aquatic environments and fisheries. This study aimed to estimate the energy fluxes in the food webs of Santa Cruz and Umari reservoirs, simulate the impacts of the Nile tilapia farming in net cages on the biotic communities and their interferences in the fishing activity; in addition to determining the carrying capacity for fish farming in both reservoirs. Two models were used: a phosphorus mass balance model to estimate the concentrations of phosphorus in the water of the reservoirs; an Ecopath with Ecosim model for estimating the energy fluxes in the food webs, ecosystem properties and simulation of the influence of the net cages on the fishing activity of each reservoir. Data were collected for three years from 2012 to 2014 in both reservoirs to calculate the biomass of aquatic communities (plankton, macrophytes, benthos, fish) and to determine nutrient concentrations in water. Quantitative information on fishing and fish farming activity was also collected. The results indicate that Umari has a higher carrying capacity than Santa Cruz, mainly by the lower emission of phosphorus to the reservoir, so that Umari could receive a larger production of tilapia in net cages. The food webs of both reservoirs were based on detritus, being the main source of energy for aquatic communities. Both reservoirs were considered immature environments with low stability, however, the accumulation of detritus in these environments may represent a energy stock to be used during disturbances. In the simulations, all aquatic communities benefited from the nutrient supply from the fish farming, increasing their populations. With the simulation of the increase of the aquaculture production it was evidenced the increase in the catches by fishing of the main commercial species