Abstract: P300-based brain-computer interface (BCI) has been widely applied due to its high accuracy and information transfer rate (ITR). Previous research showed that the intensity of visual stimuli would affect the performance of brain-computer interface systems. In this paper, we proposed a novel configuration of visual stimulus on grey value. Firstly, the RGB full range (from 0 to 255) was divided into 9 levels according to the grey-scale sensitivity of the human eye. This is for exploring the intrinsic effects brought by the grey value about the amplitude and response time (RT) of P300 wave, and the accuracy and ITR of the BCI system. In this new paradigm, 9 kinds of stimuli were displayed in the assigned squares distributed on the black (255, 255, 255) background. During this process, the stimulus was presented with random grey values in target squares or in non-target ones, which required every subject to focus his/her mind on every target flicker in the target square in spite of the differences in grey values. 18 healthy subjects were invited to do the experiment where 16 channels were utilized. It was found that the increasing in grey value produced fluctuation increasing in amplitude, accuracy and ITR, while the increasing in grey value resulted in decreasing in RT with fluctuations. Moreover, the grey value 2 showed the highest averaged offline accuracy, ITR and amplitude of P300, while the grey value 3 performed the highest averaged online accuracy and ITR. These results provided evidence for stimulus design and grey value selection.