The Ph.D. Thesis is devoted to the development of decision-making method that can be used under two sources of uncertainty – non-determinism and fuzziness. The proposed method is based on the fuzzy granule notion. The research has yielded a decision support method that includes the following stages: generation of alternatives and criteria; description of the alternatives with the help of fuzzy granules; evaluating probability that the alternatives' criteria values will be equal to the desired value; ranging of the alternatives; sensitivity analysis of the designed model and analysis of the results obtained. Existing methods for decision-making under one source and several sources of uncertainty have been analysed. During the development of the method for alternatives' informativeness evaluation, the entropy generalisation problem to the case of interval probabilities has been solved. Besides that, it is shown that the entropy obtained is additive as well. The method developed has practical applications, which is demonstrated by the examples considered. One of the examples considers project selection problem. The example analyses pulp mill construction projects based on uncertain input data. The second example considers forecasting of a product's life cycle stage. Object-oriented software has been developed in Microsoft Visual C++ 6.0 environment in order to carry out the experiments. The Thesis consists of 156 pages, 55 formulae, 45 figures and 18 tables.