In this diploma thesis, the Rijeka coast is analyzed in detail, with special emphasis on the quality of the sea in the western part of the river. Representatives of sea quality on the beaches were selected - Kantrida-east beach, Kantrida-west and Bivio. For the purposes of this paper, the Weka software system was used, which uses machine learning methods to predict the level of bacteriological contamination, Escgerichie coli and intestinal enterococci on the mentioned beaches. The city of Rijeka is located on the northern coast of the Gulf of Rijeka, and makes up most of the Kvarner Bay. The Rijeka area is especially characterized by karst springs. Such springs are caused by the process of karstification, which creates cracked systems in the underground by which water is drained, and further springs from abundant coastal springs and springs. This poses a major problem for marine quality, as often such infiltrated rainwater from urban areas in multiple relationships and various pollutants, and even microbiological loads of the coastal sea. Also, it is necessary to mention the occurrence of heavy and frequent rains in the Rijeka area. Heavy rains are often triggered by rain overflows and sewage spills into the sea, but also by the raising of sea sludge, which impairs the quality of the sea in this area. It is all of the above that has caused poor sea quality on the beaches of Kantrida-East and Kantrida-West. Since two of all three analyzed sites have a high degree of visitors during the bathing season, it is important that the quality must be good or excellent so as not to endanger their users. One of the most important procedures of quality prevention must organize supervision, in order to take adequate measures at the time of the enterprise. For the purposes of preparing the diploma thesis, the measured values of Escherichia coli and intestinal enterococci were used at the localities Bivio, Kantrida-west and Kantrida-east, in the period from 2009 to 2016. Also, the recorded values of precipitation, air temperature and wind at the postage of Rijeka, Matulji and Marčelja were used as input data. Since the data from the Matulji station gave the most adequate results, they were used in further analysis. The analysis of the sea quality for the Bivio locality showed the obtained results of weak interconnection of quality sea waters with precipitation, ie hydrological conditions on land, while for the remaining two they resulted in a stronger interconnection.