Theory of optimization which is provided in financial theory help the investors to choose better portfolios. Markowitz mean-variance model is the first method which introduces in this field. However, in experimental uses, there are deficiencies such as errors in estimates of model inputs. Due to problems in Markowitz model and in order to provide an efficient model for covering errors of previous models, Balck-Litterman asset allocation model introduced. In use of Balck-Litterman model which have the ability to combine investor information with information dived from observations, have a better estimate of the parameters for improving results of the mean-variance model.In this research in the use of Balck-Litterman model, investors considered to be non-deterministic and to show the uncertainty, fuzzy set theory is used. Then based on historical data and market specialists, Markowitz model and initial Balck-Litterman model and fuzzy Balck-Litterman model, evaluated with efficient frontier and evaluation criteria such as Sharpe, Jensen, Treynor, M 2 and appraisal ratio. The proposed model in this thesis had outperformed Markowitz model, initial Balck-Litterman model and similar availabl model.
