This paper develops an event-related potential -based adaptive model for the control of humanoid robot movements in an environment with cluster obstacles based on live video feedback. This model adaptively determines the repetition number according to an individuals’ mental state to speed up the robot control cycle. N200 and P300 potential features increase in the frontal and occipital areas when using robot images as visual stimuli, so it is able to effectively recognize target visual stimuli by processing Fisher’s linear discriminant analysis and to identify a subject’s intention by using support vector machine, in parallel. The off-line evaluations show that, compared with a non-adaptive model, the adaptive model increases the accuracy rate from 88.8% to 92.9% a change of 4.1% and the information transfer rate from 41.3 bits/min to 46.3 bits/min a change of 5.0 bits/min. Eight subjects participated in telepresence controlling a NAO humanoid robot to move in an office environment with cluster obstacles. The successful maneuvers demonstrate that the brain-controlled humanoid robot can be applied for surveillance and exploration in unknown environments based on live video feedback, which are Manuscript received October 18, 2015; revised January 19, 2016; accepted January 22, 2016. Copyright (c) 2016 IEEE. Personal use of this material is permitted. However, permission to use this material for any other purposes must be obtained from the IEEE by sending a request to firstname.lastname@example.org. This work was supported in part by the National Natural Science Foundation of China under Grant 61473207, and in part by the State Key Laboratory of Robotics at Shenyang Institute of Automation under Grant 2014-Z03. M. Li is with the School of Electrical Engineering and Automation, Tianjin University, Tianjin 300072, China (e-mail: email@example.com). W. Li is with Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China, with the School of Electrical Engineering and Automation, Tianjin University, Tianjin 300072, China, and also with the Department of Computer & Electrical Engineering and Computer Science, California State University, Bakersfield, California 93311, USA (corresponding author, e-mail: firstname.lastname@example.org). L. Niu is with the Department of Math and Computer Science, West Virginia State University, Orangeburg , SC 29118, USA (email@example.com) H. Zhou is with McGovern Institute for Brain Research, Massachusetts Institute of Technology, Cambridge, MA 02139, USA, and also with Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong 518055, China (e-mail: firstname.lastname@example.org). G. Chen is with Intelligent Fusion Technology, Germantown, MD 20876, USA (email@example.com). F. Duan is with the Department of Automation and Intelligence Science, College of Computer and Control Engineering, Nankai University, Tianjin 300071, China (e-mail: firstname.lastname@example.org). evaluated by using the new metrics for the performance of the BRI system.