Augmented reality (AR) technology using a head mounted display (HMD) is one of the fundamental tools in the next smart internet of things (IoT) society. Nowadays, portable brain-machine interfaces (BMIs) using an HMD have been studied for the future of BMI interlocked with the present IoT technology. In order to investigate the feasibility of the top-down SSVEP (steady-state visual evoked potential) BMI embedded in an HMD, SSVEP stimuli was presented in a HoloLens (Microsoft) for augmented reality (AR) constructed by holography. Electroencephalogram (EEG) was measured during the top-down SSVEP-based BMI performance, where a grid-shaped flickering visual stimulus was presented in the display of HoloLens. We examined its feasibility in a real-Time basis by its decoding accuracy. We found that the top-down SSVEP-BMI could be efficiently embedded in an AR-based HMD, and thus it can be applied for the AR-based device-control automation in an IoT space using EEG signals.