Often small and middle-size systems usually lack of a dynamical model which is not known in all. Moreover most of them are based on those classical PID controllers featured in not having a controlling system through a setpoint variable during the whole time where the process is in evolution. In this paper, a strategy of control based on a Nonlinear Model Predictive Control (NMPC) for nonlinear systems together to its respective technique of system identification, is proposed. In concrete, the approach taken for system identification is based on the approximate model of Volterra series where its respective kernels are projected onto Orthogonal Basis Functions (OBF). By using the previous model we shall test a NMPC technique which uses a similar technique to that of the Dynamic Matrix Control (DMC). The curve of control of temperature is compared to the case when a PID is used by showing the advantages of the scheme proposed. The usage of this scheme would be interesting to those facilities where the temperature trajectory tracking turns out to be crucial to obtain a high quality final product. © 2010 IEEE.