This letter investigates the artificial noise aided beamforming design for secure simultaneous wireless information and power transfer (SWIPT) in a two-tier downlink heterogeneous network, where one femtocell is overlaid with one macrocell in co-channel deployment. Each energy receiver (ER) in femtocell can be considered as a potential eaves- dropper for messages intended for information receiver (IR). Our objective is to maximize the secrecy rate at IR subject to the signal-to-interference-plus noise ratio (SINR) requirements of macro users (MUs), transmit power constraint and energy harvesting constraint. Due to the non-convexity of the formulated problem, it cannot be solved directly. Thus, we propose a novel reformulation by using first-order Taylor expansion and successive convex approximation (SCA) techniques. Furthermore, an SCA-based algorithm with low complexity is proposed to arrive at provably convergent solution. Finally, numerical results evaluate the performance of the proposed algorithm.