The Constant Modulus Algorithm (CMA), although it is the most commonly used blind equalization technique, converges very slowly. The convergence rate of the CMA is quite sensitive to the adjustment of the step size parameter used in the update equation as in the Least Mean Squares (LMS) algorithm. A novel approach in adjusting the step size of the CMA using the fuzzy logic based outer loop controller is presented in this paper. Inspired by successful works on the variable step size LMS algorithms, this work considers designing a training trajectory that it overcomes hurdles of an adaptive blind training via controlling the level of error power (LOEP) and trend of error power (TOEP) and then obtains a more robust training process for the simple CMA algorithm. The controller design involves with optimization of training speed and convergence rate using experience based linguistic rules that are generated as a part of FLC. The obtained results are compared with well-known versions of CMA; Conventional CMA, Normalized-CMA [Jones, IEEE conference record of the twenty-ninth asilomar conference on signals, systems and computers (Vol. 1, pp. 694-697), 1996], Modified-CMA [Chahed, et al., Canadian conference on electrical and computer engineering (Vol. 4, pp. 2111-2114), 2004], Soft Decision Directed-CMA (Chen, IEE Proceedings of Visual Image Signal Processing, 150, 312-320, 2003) for performance measure and validation.