This paper leverages fluid antenna (FA) and rate-splitting multiple access (RSMA) to enhance the physical layer security (PLS) of an integrated sensing and communication (ISAC) system. We consider a practical multi-user multi-input single-output (MU-MISO) system, where a base station (BS) equipped with fixed position antennas (FPAs) employs RSMA to communicate with multiple single-FA users, while an eavesdropping target may potentially wiretap the signals. The system adopts a novel rate splitting (RS) scheme, where the common layer stream serves a dual purpose: it conveys valid data to legitimate users (LUs) while simultaneously generating jamming signals to confuse potential eavesdroppers. We establish the problem and propose the optimization algorithm under two conditions: perfect and imperfect channel state information (CSI) conditions. Specifically, under perfect the CSI condition, we address the non-convex optimization problem by proposing an alternating optimization (AO) algorithm, which decomposes the problem into two subproblems: beamforming matrix optimization and the adjustment of FA positions. For beamforming optimization, we utilize semidefinite programming (SDP) and successive convex approximation (SCA) to convert the problem into a more tractable convex form. Given a fixed beamforming matrix, SCA is applied to handle the surrogate upper bound of the constraints. In the case of imperfect CSI, the continuous nature of CSI errors leads to an infinite number of constraints. To overcome this challenge, we propose an AO-based algorithm that incorporates the S-Procedure and SCA to obtain a high-quality beamforming matrix and effective FA positions. Extensive simulation results demonstrate that the proposed FA-aided RSMA-ISAC system significantly enhances security compared to traditional FPA-based and SDMA-based systems.