The QoS-Based Web services composition optimisation problem is an NP-Hard problem. So far, solving such a problem consists on finding its optimal solution, while optimizing an objective function using the QoS as an optimization criteria. In this paper, we propose an approach based on the use of a new Extended Bat Inspired Algorithm to deal with the QoS-Based Web Services composition optimization problem. The Bat Inspired Algorithm has the advantage of providing a very quick convergence at a very early stage by switching from exploration to exploitation. This makes it an efficient algorithm. The originality of the proposed approach is the designing and the built of the composition solutions by adjusting the main parameters of the algorithm. Then, to compare potential generated solutions, different QoS attributes are considered and aggregated at the complete composition level. A prototype has been realized and applied to a text translation case study. The results of experimentation are very encouraging and show that the approach is highly efficient in terms of optimality rate and running time.