The actual communication networks, due the complexity in their design and the development of new network technologies that needs better CPU performance, have needed frameworks to make projections about their growing, taking into account that they must evaluate network requirements such as bandwidth, latency and CPU performance. Kalman filter, due the capacity to bring near in dynamic systems, is a set of mathematical equations that allows to estimate in past and present tense, and make future tense projections, even if the nature of the modeled system is unknown. The analysis stage is about making projections of the variation of active cellphone lines in Colombia, using Kalman filter, to propose a new tool that meets the goal of make projections, and take preventions.