Stochastic resonance (SR) of a tristable system driven by dichotomous noise (DN) is investigated firstly by the mean signal-to-noise ratio gain (SNR-GM). Utilizing an efficiently numerical algorithm, we acquire the asymmetric DN accurately. Then the system responses and the SNR-GM as the signatures of the stochastic resonance are calculated by the fourth-order Runge–Kutta algorithm. It is founded that the change of system parameters [Formula: see text] and [Formula: see text] in a certain range can induce SR phenomenon. Moreover, with the increase of parameter [Formula: see text], the amplitude of SNR-GM increases and shows the trend of moving to the left. For the different state values of the symmetric DN, the SNR-GM will increase with the increase of state value [Formula: see text] and [Formula: see text] but only a highest peak and the interval of SR shift to the left. However, with the increase of forcing frequency, the SNR-GM decreases and the interval of SR moves to right. In addition to, the highest peak of SNR-GM will decrease with the increase of periodic signal amplitude. These results provide a reliable basis for how to realize the parameter selection of stochastic resonance in tristable system driven by DN.