Through this article the development of a system capable of recognizing the basic steps of a horse in a natural environment is shown. This development is focused on artificial intelligence, where using the processing of a PC, reference algorithms are implemented to treatment and recognition of signs of equine movements captured by inertial sensors. This process is used Fast Fourier transform and artificial neural networks in the software component, the electronic implementation includes the use of the board Enpic14® and Zig-Bee protocol for communicating portable device located on the horses and the computer. The result is a recognition system equine basic steps for identification and characterization of livestock ready for target practice mounted at the National School of Carabineros "ESCAR". This work is developed by the research group in software and Facatativá "GISTFA" technologies University of Cundinamarca in partnership with the research group of the National School of Carabineros "Alfonso Lopez" ESCAR-DINAENro.COL0061592 under the research project "Design of a simulator for shooting lessons mounted police national school" Alfonso Lopez", national police approved in 2014.