Nonorthogonal multiple access (NOMA) is a promising technology in future mobile communication systems. In this paper, considering that the base station knows imperfect channel state information (CSI), we investigate the robust beamforming design problem for NOMA systems in multiple-input-single-output (MISO) channels. Modeling channel uncertainties by the worst-case model, we aim at maximizing the worst-case achievable sum rate subject to the transmit power constraint at the base station. We propose to decouple the nonconvex optimization problem into four optimization problems and employ an alternating optimization algorithm to solve the problem. Simulation results demonstrate that our proposed robust beamforming scheme outperforms the orthogonal multiple-access scheme.