The Simon effect arises if a spatial response feature is used to represent one or more responses in working memory. The previous studies found that the go/no-go task in which stimuli were the same as the standard Simon task can generate Simon effect only when the participants used the spatial feature to represent response by referring to the location of the nonresponding hand. However, the participants can select ?exibly the reference frame for action according to sensory context. Further, the reference frame determined whether spatial anatomical code or the cognitive spatial code is used to represent the one response in the go/no-go task. Thus, we suppose that sensory context of the nonresponding hand may affect the nature of the Simon effect in the go/no-go task. If the spatial anatomical code is used to represent the response, then the visuomotor Simon effect would arise; if the cognitive spatial code is used to represent the one response, then the cognitive Simon effect would arise. To explore the supposition, we implemented three experiments. In the experiments, the stimuli were red or green squares presented in left and right physical locations, participants were instructed to perform a spatial response to one color, and refrain from responding to another color. But, the status of the nonresponding hand were different: In Experiment 1, the participants were only required to press one key by index finger of one hand and put the nonresponding hand on any place except for the desk and keyboard; in Experiment 2, the nonresponding hand was asked to put on a fixed location of the desktop; in Experiment 3, the subjects were instructed to put their nonresponding hand on a fixed key (e.g., if the response key was “Z” then put the nonresponding hand on “/”, and vice versa). Moreover, if the Simon effects were obtained in the experiments, we would further explore the different mechanism of Simon effects with reaction time (RT) distribution analysis. To that end, the correct RTs of each participant are ‘‘vincentized’’ in separate experimental condition, i.e., rank-ordered RTs are divided into equally sized quantiles, and the mean RT was grouped into equally spaced quantiles. The results showed: in Experiment 1, we didn’t obtain the Simon effect, F(1,21) = .39, p = .54; in Experiment 2, we obtained a significant Simon effect, F(1,20) = 8.48, p < .01; in Experiment 3, we also acquired a significant Simon effect, F(1,25) = 10.93, p < .01. Moreover, the RT distribution analysis showed that the size of Simon effect in Experiment 2 was rather stable from the first to fourth quantile but disappeared in the fifth quantile, and that in Experiment 3 increased from the first to the fifth quantile. The results demonstrated that representing the response by the spatial feature is the prerequisite for the generation of the Simon effect. Moreover, the results implied that the nature of the Simon effect relied on the sensory context of the nonresponding hand.