This research focuses on the georeferenced location of distributed resources, specifically the injection of active power through distributed generation. A rural-urban marginal feeder of a distribution company in Ecuador with georeferenced information was taken as a case study, which has a three-phase primary link at a medium voltage and several single-phase branches at a medium voltage of great length to supply users who are far away from the local company's network. Consequently, to analyze the behavior of the electrical network, the Cymdist software was used to perform simulations in a steady state without and with the insertion of distributed generation. For the location of the distributed generation, the voltage collapse prediction index was used as a technique for quantifying and identifying problems in the network nodes. Moreover, based on the proposed methodology, the optimal georeferencing of the sites where it is necessary to inject active power to improve the voltage profiles and reduce the voltage collapse prediction index was obtained.