Wastewater treatment plants are dynamic systems subject to wide and uncontrolled variations of flow, concentration and composition of effluents crossing-by the plant. Mathematical models are essential to describe, predict and control the variation of method such process. The aim of this work is to develop a model for monitoring the process of an activated sludge waste water treatment plant located in the suburb or Rabat, the capital of Morocco. Data of the several physic-chemical parameters (i.e. biochemical oxygen demand -BOD5; chemical oxygen demand -COD; flow; total phosphorus; total nitrogen; dissolved oxygen; suspended solids (SS , MVS ); temperature (T);and pH) were recorder over the period 2004-2008 and assessed using multidimensional data analysis methods for correlations. Results showed a significant correlation between SS and SLE and between BOD5 and COD. Moreover, a significant correlation between MVS and MES, and between BOD5 and total nitrogen were observed at the at process output. Regression of least squares partial (PLS) was performed on significant parameters and resulted in three equations, with significant correlation coefficients suggesting a good sensitivity of the built model.