The integration of aerial platforms to provide ubiq- uitous coverage and connectivity for densely deployed terrestrial networks is expected to be a reality in emerging sixth-generation networks. Energy-effificient design and secure transmission are two crucial issues for integrated terrestrial-aerial networks. With this focus, due to the potential of RIS in substantially saving power consumption and boosting the security of private information by enabling a smart radio environment, this paper investigates the energy-efficient hybrid beamforming for multi- layer reconfigurable intelligent surface (RIS)-assisted secure in- tegrated terrestrial-aerial network for defending against simul- taneous jamming and eavesdropping attacks. Specifically, with the available of angular information based imperfect channel state information (CSI), we propose a framework for the joint optimization of user’s received precoder, terrestrial BS’s and HAP’s digital precoder, and multi-layer RIS analog precoder to maximize the system energy efficiency (EE) performance. For the design of received precoder, a heuristic beamforming scheme is proposed to convert the worst-case problem into a min-max one such that a closed-form solution is derived. For the design of digital precoder, we propose an iterative sequential convex approximation approach via capitalizing the auxiliary variables and first-order Taylor series expansion. Finally, a monotonic vertex-update algorithm with penalty convex concave procedure is proposed to obtain analog precoder with low computational complexity. Numerical results show the superiority and effective- ness of proposed optimization framework and architecture