In order to improve drinking water treatment plant processes on the Butoniga reservoir in Istria (Croatia), the prediction of Water Quality Index (WQI) is done. Based on parameters such as temperature, pH, turbidity, KMnO4, NH4, Mn, Al and Fe, the calculation of WQI is conducted, while for the prediction models of WQI along with mentioned parameters, additionally O2, TOC and UV254 are used. All data are pre-processed concerning modelling and research goal. For the WQI prediction models, the entire span of the measured daily data was used, from 2011 to 2020, while missing data were filled in by usage of cubic spline interpolation. Four models were built to predict WQI with time step of one, five, ten, and fifteen days in advance in order to improve certain processes on the drinking water treatment plant regarding changes in raw water quality in the Butoniga reservoir which could influence the treatment processes. Therefore, these models can help in optimisation of certain processes which are depending on the quality of raw water. For the experiment, the machine learning algorithm M5P for rule-based models integrated in the WEKA modelling software, was used. Predicted WQI (one, five, ten, and fifteen days’ time step) was set as a target (dependant) variable, whereas water temperature, pH, turbidity, KMnO4, NH4, Mn, Al, Fe, O2 TOC and UV254 were set as independent variables (descriptors) from which the predicted WQI was modelled. The above parameters were used because they represent the parts of the system (drinking water treatment plant) on top of which the target variables depend. To achieve the highest correlation coefficient (R), and the optimal number of rules default values of parameters for building models were used in WEKA modelling software. The model that was performing most accurately, according to the validation method, was selected as a representative model for the prediction purposes. Obtained results from models showed that, one day prediction model for WQI has only one rule which is related to a linear equation predicting the WQI and very high correlation coefficient of 0.93, while five day prediction model is consisted of 12 rules with correlation coefficient of 0.81. Ten day prediction model has 14 rules and correlation coefficient of 0.79. Finally, fifteen day prediction model has five rules and correlation coefficient of 0.71. As expected, obtained correlation coefficient decreases as the number of prediction days increases, while the number of rules, and also related linear equations depends on the parameters set in WEKA modelling software which give the highest values of correlation coefficient. Also, all models have high accuracy compared to the measured data, with a good prediction of the peak values included Therefore, obtained models through the prediction of WQI could help to manage certain processes on drinking water treatment plant which are depending on the quality of raw water in Butoniga reservoir.