An oil field exploration and development campaign is bounded with different kinds of uncertainty. The most basic one that an E and P portfolio manager deals with is the one related to the existence (or not) of oil in a given prospect. Typically, technical uncertainties are related to learning, and tend to be reduced with investments on information acquisition. From the correlation pattern on the prospects in a given exploratory portfolio, follows that the results from one initial wildcat drilling will, potentially, reveal, additional information about the oil existence (or not) in other prospects in the same geological play. This way, each prospect to be drilled might be understood as a learning option to be exercised (or not) depending on its respective success probability. In such case, one of the main factors on optimizing the exploratory campaign is choosing the ideal drilling sequence. Such choice is more complex, as the quantity and diversity of the prospects increases. Given such background, the present work proposes a model that intends, using Genetic Algorithms, to optimize the drilling sequence and, as a consequence, the total portfolio value. The proposed model considers the interdependencies and each prospect specific aspects and has as an objective function (to be maximizes) the portfolio net present value (NPV). Options and learning are the main aspects underlying the optimization model. The model was evaluated on ten different exploratory portfolios and, in every case, was able to deliver at least one sequence that could represent expressive NPV gains compared to the basic scenario.