The Capacitated Arc Routing Problem (CARP) is a challenging combinatorial optimization problem with wide real world applications. Since CARP is NP-hard, many heuristic and meta-heuristic approaches have been applied to solve CARP. In this paper, a novel extension memetic approach, named MAENS-Pls is proposed to improve the state-of-the-art algorithm - Memetic Algorithm with Extended Neighbourhood Search (MAENS) for CARP, in which using an adaptive probability to optimize the frequency and depth of local search. Based on the parameter optimization on all the 8 instances, experimental studies show that the proposed algorithm can significantly improve the average performance of MAENS, and there is an increase in the robust stability and convergence reliability. Furthermore, 4 new best known solutions are found by the proposed algorithm.