Large players in energy market dedicate many efforts in valuation and optimal capital allocation decision for their project implementation, due the large candidate projects number in their investment portfolios. These decisions aim to choose the projects subset to be implemented, because the monetary resources are generally smaller than necessary for all projects implementation. There are many risks, and with risks and uncertainties, greater become the difficulties in analysis and optimally investment decisions. The classical methods to investment portfolios are based on to maximize returns (NPV, IRR, among others) and to minimize risks (NPV standard deviation, variance, among others). Often, these traditional methods may not be able to handle properly the projects managerial flexibilities (Real Options), as well the risks and uncertainties, due to possible solution difficulties and mathematical modeling problems (multi variables). Alternative models development and implementation, such as those based on Real Options Theory, including the use of Computational Intelligence methods, may be more suitable for these problems. In this thesis, a hybrid methodology is developed, presenting a Fuzzy Real Options model for Revamp projects valuation by a Thermoelectric Power Generation market player, from a Real Options Portfolio in uncertainties environment. For selecting the projects subset by budget range, a multi-criteria Genetic Algorithm optimization is applied, using a weighting return x risk index (lambda).