Based on the “solution-diffusion” theory, the developed model in this study is able to determine the permeate water quality and the specific energy consumption of a reverse osmosis system. Python programming language is chosen due to its interpreted, high-level, general-purpose nature and its emphasis on code readability. The results show a good fit with two commercial software: Lewaplus and ROSA. This model also takes advantage of the PHREEQC chemical calculation engine for solution analysis and scaling prediction. Because saline groundwater often contains ions such as Ca2+, Mg2+ and CO32-, a scaling investigation with different indices is needed for a stable operation of RO system. Further, the ion selectivity of three different types of membrane (BWRO, SWRO and NF) are implemented in the model via the corresponding rejection parameters. Two performance criteria of the model are the permeate water quality and the transmembrane pressure which are used in the calibration and validation. The calibration process uses NaCl solution with various operating condition to determine solvent and solute permeability of a membrane. The calibration procedure can be conducted with either experimental, literature or simulated data. The validation results with experimental data are not ideal due to dissimilarity of active membrane area, layer structure (flat sheet vs spiral wound) and recovery rate. The result from simulation of different type of water samples proves that selection of membrane type should be revolve around the quality of feed water. Three suggestions on system design are offered regarding the water samples from China, Oman and Egypt.