Discharge of brine wastewater produced from industrial plants into adjacent coastal water bodies is considered as a preferable and common method currently used in many offshore industrial plants. Therefore, it is important to carefully study the behavior of jets and their environmental impacts on water bodies close to the discharge points, especially when the density is different between the jets and the receiving water. The main goal of this study is to improve the understanding of the mixing behaviour of jet trajectories for positively (offset) and negatively (inclined) buoyant jets when density is considered a significant factor, and also to examine the accuracy of some RANS turbulence models and one type of artificial neural network in predicting jet trajectory behaviours. In the first part of this study, experiments using a PIV system for offset buoyant jets were conducted in order to study the effect of the density differences (due to salinity [nonthermal] or temperature [thermal]) between the discharge and the receiving water body on the jet behavior, and the results showed that the nonthermal jets behaved differently as compared to the thermal jets, even though the densimetric Froude numbers (Frd) and density differences (∆ρ) were similar. In addition, a Reynolds-averaged Navier-Stokes (RANS) numerical model was performed using open-source CFD code (OpenFOAM) with a developed solver (modified form of the pisoFoam solver). The realizable k-ε model showed the best prediction among the models. Secondly, an extensive experimental study of an inclined dense jet for two angles (15°and 52°) was conducted to study the effect of these angles on the jets’ geometrical characteristics in the presence of a wide range of densimetric Froude numbers as well as with different discharge densities. More experimental data were obtained for these angles to be added to the previous data for the purpose of calibrating, validating, and comparing the various numerical models for future studies. The results of these experiments are used to evaluate the performance of a type of artificial neural network method called the group method of data handling (GMDH), and the GMDH results are then compared with existing analytical solutions in order to prove the accuracy of the GMDH method in simulating mixing behaviors in water bodies. Thirdly, a comprehensive study on predicting the geometrical characteristics of inclined negatively-buoyant jests using GMDH approach was conducted. The superiority of this model was demonstrated statistically by comparing to several previous analytical models. The results obtained from this study confirm that the GMDH model was highly accurate and was the best among others for predicting the geometrical characteristics of inclined negatively-buoyant jests.