<div>In this paper, the global exponential stabilization of two-level stochastic quantum systems is achieved by using measurement-based feedback strategies, including state feedback and noise-assisted feedback, and the state convergence rate is improved by dividing the state space. For both the combination of two continuous noise-assisted feedback and the combination of state feedback and noise-assisted feedback, the state space is divided based on the comparison of real-time state convergence rate under different measurement-based feedback strategies, respectively. The global exponential convergence and the effect of dividing state space in improving state convergence rate for two-level quantum systems are proved in theory and verified in numerical simulations.</div>