This paper contains the basic features of the Plitvice Lakes, hydrogeology and meteorological characteristics of the area, and the basic characteristics of the lakes. The analyses of hydrological parameters of the lake and flow have been made at an annual, monthly and daily basis. In addition to the aformentioned analysis, there has been made an analysis of the parameters of water quality at stations Bijela and Crna river, as well as Matica and Kozjak lake. The paper describes the methodological approach and the results of modeling the daily water level and flow of the lake Prošce and Kozjak. Modeling was performed using methods of machine learning. The chosen models for this thesis are neural networks with multilayer perceptron and regression decision trees in the software package Weka. Modeling of hydrological parameters was carried out as a function of hydrological parameters of neighboring stations, and meteorological characteristics of the area. By means of modeling the parameters of water quality were obtained, as a function of other parameters of quality, and hydrological and meteorological characteristics to boot. The quality of the results of modeling has been evaluated by the correlation coefficient and the values of the mean absolute error. Finally, the modeling attributes of water quality has been made on the river Matica in the function of hydrology and water quality parameters of the Bijela and Crna river.