In this paper, a novel reinforcement learning adaptive risk-sensitive fault-tolerant stochastic non-affine integrated guidance and control (NAIGC) method is proposed for a class of skid-to-turn (STT) missiles. The cost standard of the risk-sensitive index we coveted can be arbitrarily small, ensured by solving a specific inequality. Firstly, an extended integration system is introduced to solve the challenging control problem posed by the non-affine form of the control signal, taking into account the non-affine nature in the missile. Secondly, for random noise and unknown non-linearities in NAIGC systems, a reinforcement learning actor-critic adaptive risk-sensitive control method is proposed to ensure the input-state stability of the system. Subsequently, hyperbolic tangent functions and adaptive boundary estimation are used to reduce disturbance-induced jitter and actuator fault-induced bias in the control system. In addition, the proposed control strategy improves the missile's interception performance against maneuvering targets and reduces the conservatism of existing adaptive robust control methods. Ultimately, not only is the stability of the NAIGC closed-loop system demonstrated using Lyapunov theory, but the effectiveness and superiority of the method is also verified through numerical simulations.