The construction of an oil well requires extensive and advanced planning. Among the various objectives of this planning is the verification of the need to purchase materials and equipment to carry out the stages of construction of the well. Such acquisitions often involve long contracts and, later, require a long lead-time, which can reach years. As this planning is carried out in a scenario of many uncertainties, several techniques, using different types of data, have already been proposed to correlate wells, in order to obtain in advance the material and equipment requirements to build a new well. One of these types of data is the lithological profile, which contains the rock segments present throughout the length of the well, collected through sensors and other methods present during the drilling. It is possible to generate artificial lithological profiles for not yet drilled wells, through seismic data. This work proposes a new methodology for grouping oil wells. The distance measure is based on the degree of similarity between wells, obtained by applying a sequence alignment algorithm, which, in turn, are generated exclusively from the lithological profiles of such wells. In this way, it is possible to obtain wells related to a specific well. To validate the methodology, clustering experiments involves data from 120 wells on the southeastern Brazilian coast.