In the Gaza Strip, the available freshwater sources are severely polluted and overused. Desalination of seawater through reverse osmosis (RO) has become the most realistic option to meet a rapidly growing water demand. It is estimated that the Gaza Strip will need to develop a seawater desalination capacity of about 120,000 m(3)/d by the year 2008, and an additional 30,000 m(3)/d by the year 2016 in order to maintain a fresh water balance in the coastal aquifer and to fulfill the water demand for different uses in a sustainable manner. Cost and reliability of a large RO facility are still subject to much uncertainty. The cost of seawater desalination by RO systems varies with facility size and lifetime, financing conditions, intake type and pre-treatment requirements, power requirements, recovery rate, chemicals cost, spare parts cost, and membrane replacement cost. The permeate salinity is a function of feed water temperature, recovery rate, and permeate flux. The quantity of water produced depends mainly on plant size, recovery rate, and operating load factor. Many of these parameters are subject to a great deal of uncertainty. The objective of this work is to develop a probabilistic model for the simulation of seawater reverse osmosis processes using a Bayesian belief network (BBN) approach. This model represents a new application of probabilistic modeling tools to a large-scale complex system. The model is used to: (1) characterize the different uncertainties involved in the RO process; (2) optimize the RO process reliability and cost; and (3) study how uncertainty in unit capital cost, unit operation and maintenance (O&M) cost, and permeate quality is related to different input variables. The model utilizes information from journal articles, books, expert opinions, and technical reports related to the study area, and can be used to support operators and decision makers in the design of RO systems and formulation of operational policies. The structure of the model is not specific to the Gaza Strip and can be easily populated with data from any large-scale RO plant in any part of the world.