Electroencephalogram (EEG) is one of the low-cost Brain Computer Interface (BBA) methods that understands the emotional and cognitive activity of the users and maps them to a meaningful signal. In the EEG method, which is commonly used for emotion recognition, the number of channels may affect the emotion recognition performance. In this work, the performance of different number of EEG channels on emotion recognition has been explored, in detail. Experimental findings show that a better emotion classification performance can be achieved with fewer channels.