In this paper, we reveal the branching structure for a non-homogeneous random walk with bounded jumps. The ladder time $T_1,$ the first hitting time of $[1,infty)$ by the walk starting from $0,$ could be expressed in terms of a non-homogeneous multit
ype branching process. As an application of the branching structure, we prove a law of large numbers of random walk in random environment with bounded jumps and specify the explicit invariant density for the Markov chain of ``the environment viewed from the particle .The invariant density and the limit velocity could be expressed explicitly in terms of the environment.
By decomposing the random walk path, we construct a multitype branching process with immigration in random environment for corresponding random walk with bounded jumps in random environment. Then we give two applications of the branching structure. F
irstly, we specify the explicit invariant density by a method different with the one used in Bremont [3] and reprove the law of large numbers of the random walk by a method known as the environment viewed from particles. Secondly, the branching structure enables us to prove a stable limit law, generalizing the result of Kesten-Kozlov-Spitzer [11] for the nearest random walk in random environment. As a byproduct, we also prove that the total population of a multitype branching process in random environment with immigration before the first regeneration belongs to the domain of attraction of some kappa -stable law.
