Abstract: This paper presents a pathwisely jump-adapted approximation of the Poisson driven Markov processes governed by stochastic differential equations.Any trajectory of the processes is divided into continuous phases. Within each phase, the corresponding ODEs are established and solved by the Runge Kutta schemes.The method is applied to investigate the Langevin equation as well as the Duffing-Van der Poloscillator.