Deep neural network with rectified linear units (ReLU) is getting more and more popular recently. However, the derivatives of the function represented by a ReLU network are not continuous, which limit the usage of ReLU network to situations only when smoothness is not required. In this paper, we construct deep neural networks with rectified power units (RePU), which can give better approximations for smooth functions. Optimal algorithms are proposed to explicitly build neural networks with sparsely connected RePUs, which we call PowerNets, to represent polynomials with no approximation error. For general smooth functions, we first project the function to their polynomial approximations, then use the proposed algorithms to construct corresponding PowerNets. Thus, the error of best polynomial approximation provides an upper bound of the best RePU network approximation error. For smooth functions in higher dimensional Sobolev spaces, we use fast spectral transforms for tensorproduct grid and sparse grid discretization to get polynomial approximations. Our constructive algorithms show clearly a close connection between spectral methods and deep neural networks: a PowerNet with n layers can exactly represent polynomials up to degree s n , where s is the power of RePUs. The proposed PowerNets have potential applications in the situations where high-accuracy is desired or smoothness is required. the curse of dimensionality for an important class of problems corresponding to compositional functions. In the general function approximation aspect, it has been proved by Yarotsky [12] that DNNs using rectified linear units (abbr. ReLU, a non-smooth activation function defined as σ 1 (x) := max{0, x}) need at most O (ε d k (log |ε| + 1)) units and nonzero weights to approximation functions in Sobolev space W k,∞ ([−1, 1] d ) within ε error. This is similar to the results of shallow networks with one hidden layer of C ∞ activation units, but only optimal up to a O (log |ε|) factor. Similar results for approximating functions in W k,p ([−1, 1] d ) with p < ∞ using ReLU DNNs are given by Petersen and Voigtlaender[13]. The significance of the works by Yarotsky [12] and Peterson and Voigtlaender [13] is that by using a very simple rectified nonlinearity, DNNs can obtain high order approximation property. Shallow networks do not hold such a good property. Other works show ReLU DNNs have high-order approximation property include the work by E and Wang [14] and the recent work by Opschoor et al. [15], the latter one relates ReLU DNNs to high-order finite element methods.A basic fact used in the error estimate given in [12] and [13] is that x 2 , x y can be approximated by a ReLU network with O (log |ε|) layers. To remove this approximation error and the extra factor O (log |ε|) in the size of neural networks, we proposed to use rectified power units (RePU) to construct exact neural network representations of polynomials [16]. The RePU function is defined aswhere s is a non-negative integer. When s = 1, we have the Heaviside step func...