The identification of galaxy systems is a challenging task and, therefore, it is necessary to improve these techniques as much as possible to achieve a better understanding of the structure of the universe on a large scale and link the properties of galaxies with those of the inhabiting halos. Given the large present and forthcoming galaxy catalogues, we propose, implement and evaluate an algorithm that combines the two most popular techniques to identify galaxy systems that can be easily applied to any spectroscopic galaxy catalogue. Assuming that a group of galaxies is a gravitationally bounded system that has at least one bright galaxy, we begin by identifying groups with a Friend-Of-Friend algorithm adapted to fit this definition. Knowing the shortcomings of this method, we improve it by adding a halos-based procedure. To assess the performance of the algorithm, we construct a mock catalogue from a semi-analytical model to compare the groups identified using our method with those obtained from the simulation.