Prediction of satellite positioning errors represents a substantial step towards the Global Navigation Satellite System (GNSS) performance assessment. Satellite positioning accuracy in the particular area can be expected to be similar due to prevailing environmental conditions. This similarity opens the opportunity to estimate and predict the positioning errors of close locations. The paper aims to develop a regional model of positioning errors estimation for Global Positioning System (GPS) single-frequency receivers based on ground truth data from reference stations, in this phase considering different levels of space weather activity as one f the criteria defining environmental conditions. The model should provide a simple positioning error prediction in cases where reference stations and respective data do not exist. The space weather conditions were examined to determine the influence on GPS satellite positioning performance at three selected International GNSS Service (IGS) stations in the Adriatic Region - Graz, Padua, and Matera. The mutual relations in terms of positioning error patterns were elaborated. The same 15-day period in three consecutive years was analysed. Pearson’s coefficient was utilised as a major indicator for determining the degree of correlation. The data from IGS stations Padua and Graz showed better, significant correlation results. The IGS station Matera, located farther and southward slightly differed in positioning deviations’ patterns and was not used for the model development. Satellite positioning errors of IGS Padua were used as a reference to determine the positioning errors of IGS Graz. Due to the significant correlation results, the linear regression model has been developed for the latitude, longitude, and height positioning errors. The final model coefficients were calculated as average values of the model coefficients for latitude, longitude, and height errors for elaborated periods. The cross-validation with five folds has been carried out, showing good model performance with R2 values of 0.7785 for geographic latitude, 0.8132 for the geographic longitude, and 0.7796 for height above sea level, respectively. The validation showed that the model could be applied during all levels of space weather activity on a regional basis.