This paper proposes a novel method for Blindly Separating unobservable independent component (IC) Signals (BSS) based on the use of a maximum entropy guide (MEG). The paper also includes a formal proof on the convergence of the proposed algorithm using the guiding operator, a new concept in the genetic algorithm (GA) scenario. The Guiding GA (GGA) presented in this work, is able to extract IC with faster rate than the previous ICA algorithms, based on maximum entropy contrast functions, as input space dimension increases. It shows significant accuracy and robustness than the previous approaches in any case.