Abstract: A robust beamformer design was proposed in simultaneous wireless information and power transfer systems, where a practical non-linear energy harvester circuit model was taken into consideration. In order to improve the fairness among users in the system, the minimum signal-to-interference-plus-noise ratio was optimized, thus the overall receiving performance of users could be enhanced. A non-convex optimization was formulated under the consideration of imperfect channel state information, where the transmit beamformers of information receiver were optimized. To tackle the non-convexity and the infinite constraints brought by channel estimation errors, the problem was first reformulated based on S-procedure, and further convexified by using semidefinite relaxation. An iterative beamformer design was then proposed based on bisection method. Simulation results show that the proposed algorithm can achieve 25% to 50% performance gain, compared with the non-robust algorithm. Therefore, the proposed algorithm can be applied in practical systems with non-ideal channel environment. The results also show that the proposed algorithm can achieve better fairness performance, compared with traditional beamforming such as linear energy harvesting based method or zero forcing.