Distributed Generation (DG) has deeply changed the classical structure of electrical power systems. In Brazil, the tariff model adopted by the current regulation, called the Energy Compensation System, establishes that the energy injected by DG into the distribution grid fully discounts the energy consumed, with all its tariff components. In vogue, the discussion for new rules regarding the forms of remuneration and valuation of DG involves, on the one hand, distributors, who claim that the current compensation mechanism does not adequately remunerate the use of the distribution system, and, on the other hand, consumers who choose for DG, point out its benefits to society and defend the current model. In this sense, this doctoral thesis proposes a new method for quantifying and allocating DG benefits and costs in distribution systems, which generates a tariff signal capable of translating the effects of its location and presence in the network. The basic concept covers the identification, accounting and sharing of what are called in the proposed method as functions, which represent chargeable characteristics and considered adequate in the study of DG remuneration. In this work, the functions that will be evaluated represent the DG impact on the usage, losses, peak load and reliability of the distribution network. The allocation between generators is performed using the Shapley Value of Cooperative Game Theory. In order to apply the proposed method, two test systems and a real distribution system are used and the obtained results are widely discussed.