Iran is a developing country. So, beside of developing, if we do not consider the environmental problems and specially water quality problems, economical developing will not obtain and may cause difficulties and irreparable consequences. Zayandehrood river is the main river at the central part of Iran plateau. This river has been received some wastewater and effluents from surrounding villages and several small and large industries and also from drainage systems of irrigation lands. In last decade, Zayandehrood river basin had a drought period, hence industry developing and destructive effects of drought in Zayandehrood basin which affect surface water quality and river environment, also requisiteness of possessing management model for managing quality of this river. It seems important to perform investigation about pollution in terms of quality and quantity as a management tools such as mathematical and computerized model model (blackbox model) for prediction of water quality status of Zayandehrood river and for prevention of rising the pollution. This research follows the prediction and simulation of electrical cunductivity as a water quality parameter and assessment of this parameter in drought condition. In this research, relation between electrical conductivity and hydrological parameter of Zayandehrood river is obtained by artificial neural networks as a modeling tools. Modeling tools that used in this research were MLP and RBF, finally performances of these two models were compared. For upstream of Zayandehrood river, the first three stations RBF model shows an acceptable results for modeling of electrical conductivity and at the downstream of the river, five last stations MLP had a good results. Results show that the artificial neural network can use for modeling of relation between electrical conductivity and hydrological parameters in drought conditions. Key Words: 1- Drought 2- Electrical conductivity 3- Artificial neural network 4- Multi layer perceptro 5- Radial basis function 6- Surface water quality
