This paper deal with the estimation the rotor resistance present in the dynamics of an induction motor squirrel cage using LSSVM in decline. The problem is that the rotor resistance is difficult to me asure what is required for estimating it. In addition to being affected by the increase in temperature, leading to the rotor time constant of the change and affect the engine parameters. We propose a model of the engine in Matlab Simulink, which is extracted the data, is preprocessed and applied the algorithm LSSVM in nonlinear regression with RBF kernel and gamma optimization is done by cross - validation. Performance measurement model used the square root of the average prediction error. Getting good results.