Locked-in syndrome (LIS) has to be aware of, even though it is rare in the world. However, this disease has been reported to have ten years survival rates as high as 80 %. The physical appearance of the syndrome is a total paralysis but with preservation of consciousness. Consciousness indicates that the patient's brain is still able to work quite well. Brain Computer Interfaces (BCI) is a technology that allow users to directly have control on computers or external devices by focusing on which action the users want to do through their brainwaves, which are recorded by the means of electroencephalograph. In this paper, we combine a feature of SDK community edition from a commercial BCI device called Emotiv EPOC+, which is the most credible and cost-effective mobile EEG headset on the market, with UNITY3D, the world's leading content-creation engine, to build a beautiful and engaging content in Visual Request System to enable communication between patients with LIS and the people around them. In Visual Request System, there are 2 categories of requests, which are internal request system and external request system. Both requests depend on where the request came from. We tested the system and EPOC Control Panel on 10 volunteers to get the success rate data, and we make questionnaire for comparing user-friendly interface between a Visual Request System and Cognitiv Suite in the EPOC Control Panel software from Emotiv and we also ask them about the comfort when using Emotiv EPOC+. Then, we record every calibration data of Cognitiv suite EPOC Control Panel in Xavier TestBench 3.1.18, after that we analyze what happen in calibration data using Power Spectral Density with time window in 10 seconds duration. The results of our research conclude that the Visual Request System built on UNITY3D and Emotiv EPOC+ can be used as a communication system to help LIS patients by making some improvements and the Visual Request System is more user-friendly than EPOC Control Panel to show communication message. Power Spectral Density with time window in 10 second duration analysis indicates that the signal received by the Emotiv only based on an 8 seconds calibration is intersected to one another signal. Therefore, increasing the number of mental states that must be detected by the system will make the system more difficult to distinguish each mental state signal.