The accurate estimation of financial time series provides both investment opportunities and hedging financial assets to those who take positions in financial markets. In this context, it is extremely important to estimate the price of gold, which is an important investment instrument. The research aims to estimate gold prices with various input variables the day before by using artificial neural networks method. In the study, gold prices between 11.03.2014-10.31.2019 are estimated by multi-layer artificial neural networks using variables of Brent oil prices, VIX index, Dow Jones Index and US Dollar index. As a result of the research, there is a model that best estimates (98,44%) gold prices.