This study focuses on a demonstration of differences between movement-related cortical potentials (MRCPs) during emergency and non-emergency situations. Two paradigms were designed for emergency and non-emergency situations. The necessary pre-processing and Laplacian spatial filter were used in the collected data. Then initial negative phase of MRCPs was extracted from scalp electroencephalogram (EEG) in non-emergency situation and compared with that in emergency situation. Based on the data of non-emergency, a matched filter (MF) algorithm was designed and was used to detect the motor intention in two paradigms. The result shows a significant difference of the initial negative phase of the MRCP in two cases. In addition, if the MF algorithm based on non-emergency situation was used for emergency situations directly, there was a large difference in accuracy. The true positive rate was 60.57±14.79% in nonemergency while 44.29±5.73% in emergency. The result indicates that additional consideration should be given to emergency situation when designing algorithms or collecting data. So, we designed a new algorithm to solve this problem, which works better compared to simple MF. The algorithm effectively improves the true positive rate and reduces the false positive per minute.