Based on the kinetic model for obtaining emergent bistability proposed by Tan et al. (2009), the effects of the fluctuations of protein synthesis rate and maximum dilution rate, the cross-correlation between two noises, and the time delay and the strength of the feedback loop in the synthetic gene circuit have been investigated through theoretical analysis and numerical simulation. Our results show that: (i) the fluctuations of protein synthesis rate and maximum dilution rate enhance the emergent bimodality of the probability distribution phenomenon, while the cross-correlation between two noises(λ), the time delay(τ) and the strength of the feedback loop(K) cause it to disappear; and (ii) the mean first passage time(MFPT) as functions of the noise strengths exhibits a maximum, this maximum is called noise-delayed switching (NDS) of the high concentration state. The NDS phenomenon shows that the noise can modify the stability of a metastable system in a counterintuitive way, the system remains in the metastable state for a longer time compared to the deterministic case. And the τ and the K enhances the stability of the ON state. The physical mechanisms for the switch between the ON and OFF states can be explained from the point of view of the effective potential. © 2017 Elsevier B.V.