In this letter we propose realistic schemes to realize topologically nontrivial Floquet states by shaking optical lattices, using both one-dimension lattice and two-dimensional honeycomb lattice as examples. The topological phase in the two-dimensional model exhibits quantum anomalous Hall effect. The transition between topological trivial and nontrivial states can be easily controlled by shaking frequency and amplitude. Our schemes have two major advantages. First, both the static Hamiltonian and the shaking scheme are sufficiently simple to implement. Secondly, it requires relatively small shaking amplitude and therefore heating can be minimized. These two advantages make our scheme much more practical.Topological state of matters such as quantum Hall effect and topological insulator have been extensively studied in equilibrium systems. Recently, topological classification of quantum states in a periodically driven nonequilibrium system has been proposed [1,2], in which the topologically nontrivial states are named as "Floquet topological insulator" [1]. Floquet topological band has been first realized in photonic crystal and the edge state of light has been observed [3]. While so far it has not been realized in any solid-state or cold-atom system.Realizing and studying topological state of matter is also one of the major treads for cold atom physics nowadays, for which Raman laser coupling [4][5][6][7] and shaking optical lattice [8][9][10] have been developed as two major schemes. In several recent experiments, it has been demonstrated that fast shaking optical lattices can generate synthetic abelian gauge field and magnetic flux [8,9]. In this letter we propose that shaking optical lattice is also a powerful tool to realize Floquet topological state in cold atom systems.We first demonstrate that in one-dimension lattice it realizes a system equivalent to Su-Schrieffer-Heeger model [11] with nonzero Zak phase; Then, we show that in two-dimension honeycomb lattice [12] it realizes a system equivalent to the Haldane model which exhibits quantum anomalous Hall effect [13]. So far, quantum anomalous Hall effect has only been found in chromium-doped (Bi,Sb) 2 Te 3 , and growing this material is extremely challenging [14]. It is therefore highly desirable that one can quantum simulate this effect with cold atom system. However, despite of several proposals [15] this effect has not yet been successfully simulated in cold atom setup. Our scheme has two major advantages for which it becomes much practical.The first is its simplicity. To realize a topological state in a static system, it usually requires particular form of hopping term. For instance, in order to realize the Haldane model [13], one needs to generate a special nextnearest range hopping term, which usually requires engineering laser-assisted tunneling in cold atom system [4][5][6][7]. In contrast, in our scheme, the static Hamiltonian is quite simple (it only contains normal nearest neighboring hopping without extra phase factor) and has been realized in...