This work has been focused in the application of Gustafson-Kessel Algorithm in a complex system through a methodology proposed. The complex system here considered will be the financial market. So, the main objective of this paper is to classify objects in two patterns: winner and loser. The methodology is based on application of a method of clustering called Modified Gustafson-Kessel (MGK) in some open companies of the transportation sector and energy sector. Results shows that the use of MGK can better separate the promising actions from the non-promising ones with more precision due to its covariance matrix that can be change for generate the best separability among clusters. This produces a new tool for analysis of the dynamic of stock market with the main aim of given support to investor in make decision.