Propellers are the most efficient way to generate propulsion. About 84% of the energy generated by the engines is utilized, being therefore widely used in several different aircraft. However, studies show that propellers with a diameter less than 16 inches have efficiency reduced by up to 15% when compared to larger geometries size. This deficiency is justifed for the application of mathematical models that are not so accurate for that magnitude. The present study aimed to minimize this difference increases the accuracy of the simulations performed by the Qprop software, which uses an extension of the formula with the blade element/vortex to predict the performance of different motor-propeller assemblies. For this purpose, neural networks were created that correct the results of Qprop in relation to tests carried out in wind turbines for traction and torque. For this, 28 propellers from different manufacturers and geometries were tested in wind tunnel and Qprop was used to simulate in a computational environment the same models under the same operational conditions. For the training of neural networks, geometric data of propellers and operational conditions were used as inputs. Targets are the difference between the results of the test in a wind tunnel and the Qprop simulation. To compare the accuracy of the Qprop's results and the correction made by the neural networks in relation to the tests in the wind unit, the Mean Squared Error was calculated, which came to a reduction of 99% in the case of thrust and 95% in the case of torque.