Routing is an impact factor for wireless sensor networks (WSN) longevity; in applications like precision agriculture is essential to preserve the sensor nodes lifetime because deployment field (hilly land) location usually does not allow to recharge their batteries. In this paper a genetic algorithm adaptation in order to generate routing trees for WSN in the con- text of precision agriculture is proposed. The adapta- tion consists of including in the algorithm parameters from such a context, and of determining by means of experimentation the selection method that is best suited to the solution according to the convergence time of the algorithm.