El Niño connects globally climate, ecosystems and socio-economic activities. Since 1980 this event has been tried to be predicted, but until now the statistical and dynamical models are insufficient. Thus, the objective of the present work was to explore using an autoregressive moving average model the effect of El Niño over the sea surface temperature (TSM) off the Peruvian coast. The work involved 5 stages: Identification, estimation, diagnostic checking, forecasting and validation. Simple and partial autocorrelation functions (FAC and FACP) were used to identify and reformulate the orders of the model parameters, as well as Akaike information criterium (AIC) and Schwarz criterium (SC) for the selection of the best models during the diagnostic checking. Among the main results the models ARIMA(12,0,11) were proposed, which simulated monthly conditions in agreement with the observed conditions off the Peruvian coast: Cold conditions at the end of 2004, and neutral conditions at the beginning of 2005.