This work investigates the problem of optimization of alternatives for petroleum fields` development. A development alternative refers to the way a well-known and delimited petroleum field is placed in production. This process involves the determination of the number, localization and scheduling of producer and injector wells. Thus, the optimization of alternatives consists of finding the production configurations that, in the long term, provide the maximum net present value (NPV); this is obtained from the investment cost, oil price, oil & gas production, operation costs and taxes and royalties paid during the production time. The oil and gas production is obtained from a reservoir simulator. The simulator receives information from the alternative to be simulated, and returns an oil & gas production to specified production time. Each simulation can take from a few seconds to several hours, depending on complexity of the reservoir being modeled. This work proposes, implements and evaluates an intelligent optimization system that comprises: genetic algorithms (GAs) to search an optimal development alternative; using of parallel computing environment to reservoir simulation and NPV computing; an inference module, basis in intelligent models, to approximate the oil production function; and a oilfield characterization module, basis in quality maps, to obtain information about the oilfield to use during optimization process. This work consisted of four stages: a literature review about petroleum field development and reservoir simulation; a study about computational intelligence techniques applied in optimization and functions approximation; the development of alternatives optimization proposal model; and the case studies. The proposal model was evaluated using homogeneous and heterogeneous reservoir configurations, obtaining results of optimization, by using characterization, the inference module and the parallel environment. The obtained results indicate that the proposed model provides alternatives with high NPV without previous knowledge and also from information provided by characterization or information inserted by the expert as initial seeds into optimization. The main contribution of this work is the conception and the implementation of a system basis in intelligent techniques to optimize development alternatives offering a reduction time to an iterative process, obtained from exploit of computational effort of a parallel computing environment or by using of production curves approximations. This intelligent system offers a decision-support tool that allows automating the search process of development alternatives and exploiting information from knowledge of reservoir engineers.