Throughout the history man has considered gold as a precious metal and consequently its price forecast has been of importance to him. In the past, not many intelligent, innovative approaches to price forecast were applied, and traditional methods of forecast, such as Regression, ARIMA, Exponential Smoothing, Moving Average, and methods of this kind were extensively used. Only recently Artificial Intelligence, Neural Networks and Fuzzy Logic have been proposed as forecast models. In this paper, after considering the role of gold in international finance, its demand and supply, and the relationship between gold and Dollar ,factors affecting its price fluctuations are considered and a Neuro-Fuzzy approach based on the Takagi-Sugeno Model is employed to forecast gold price. The results obtained by this method are compared with Regression Analysis, which show that a Neuro-Fuzzy yield a better and a more promising forecast.