Abstract In this work we study asymptotic properties of a long range memory random walk known as elephant walk. First prove recurrence and positive for the Then, establish transience regime model. Finally, under Poisson hypothesis, replica mean field limit obtain an upper bound expected distance walker from origin.