This paper deals with modeling lifespan and aging insulation systems in electrical machines. The different aging processes involved, including partial discharges, are described. According to the state of the art, this paper is more focused on lifespan than aging. Previous works are described and lead to the conclusion that there is no comprehensive method, even with accelerated lifetime tests, able to provide lifetime models taking into account all the stress factors in a simple way. Consequently, the principles of different regression methods are described (design of experiments, response surface, multi-linear and robust regression) and applied with a limited number of experiments, and consequently, a reduced experimental cost. Several different types of insulation materials have been tested, putting in evidence the performance of the proposed approach.