The uncertainty of the future has frightened mankind throughout history, and mankind has used many techniques in response to this in order to reduce uncertainty or to remove it. It is well known that macroeconomics, biology, medicine, engineering and social sciences are very important as well as the way of forecasting future events and conditions is in business. Preparing for the script that will take place at an unprecedented time, making plans and determining policies and ultimately making decisions will only be possible with good predictions of the future. A good estimate of this will also reduce the anxiety of uncertainty. Today, there are many estimation techniques used, such as regression analysis, time series analyzes and heuristic methods. However, each method produces different results because its infrastructure and algorithm are different from each other. Those who are directly interested in forecasting results want to know the most accurate end result analysis technique. Because, rightly, the accurate prediction of future events will provide an advantage in the intense competition environment. In this thesis study prepared by moving from point to point, different estimation techniques are combined on a set of fuzzy soft sets and a prediction value is obtained with a single output. In the analysis, the actual data on variables (inflation rate, exchange rate, gold prices, deposit interest rate, money supply, domestic debt stock, curent account balance, credit volume, oil prices, foreign trade balance and industrial production index) that are thought to have an impact on the XU 100 levels are used. The generated data set was analyzed with a univariate or multivariate estimate analysis (ARIMA, multivariate regression analysis, artificial neural networks and exponential smoothing) and the obtained results were combined on a fuzzy set. The success of each method and combination model was measured by error terms and this measurement showed that the most successful model is the combined model.