Water quality management is the most significant environmental activity today. Water quality indicators were estimated according to classification, where the integration of water quality parameters and Water Quality Regulations were taken into consideration. In this paper, the quality of primary treated urban wastewater, discharged at three locations into the Adriatic Sea, is described based on the results of measured parameters obtained daily monitoring on drain locations wider Split area over a period of 4 years. Experimental data were analyzed by basic statistical methods for determination of mean and median values, standard deviations, minimal and maximal value of the measured parameters and their mutual correlation coefficients. The results obtained showed extremely high concentrations for chlorides, sulphates, total suspended solids and for chemical oxygen demand (COD) which several times exceed the maximum allowed concentration (MAC). Median values were lower for all pollutants from their mean values except biological oxygen demand (BOD5) and pH. The highest value of the coefficient of variation was obtained for N-NO2 at Locations 2 and 3, and for detergents at Location 1. The lowest coefficient of variation was recorded for pH value at all locations. Different chemometric techniques such as cluster analysis (CA), factor analysis (FA) and principal component analysis (PCA) were established to investigate spatial and temporal characterization of pollution indicators. Spatial cluster analysis showed the distribution of clusters in two groups; cluster I is formed by Locations 2 and 3 and cluster II is formed by Location 1. The observed time period can be divided in the period 2006-2007 and the period 2008-2009 according temporal classification. The statistical analyses were used as bases for two linear models. Multivariate linear regression model (MLR) and piecewise linear regression (PLR) were made to predict the most important pollution indicators COD and BOD5 with 9 independent variables. The obtained regression coefficients for multivariate linear model amounts R2 = 0.306 – 0.448, while regression coefficients for piecewise linear regression model amounts R2 = 0.534 – 0.683. Municipal wastewater quality index in the city of Split, based on the fuzzy set rules that represent water quality via linguistic terms. After primary treatment, water quality, at Location 1 was estimated as “good” with the highest degree of membership of 0.476, while at other two locations regarding to highest degrees of membership (0.981 and 0.351), water quality was estimated as “fair”.