This study aimed to evaluate the feasibility of a video-based technology for assessing the reaction time (RT) during karate kumite situations. The Simple RT (one stimulus and one response), 2-Choice RT (two stimuli and two responses), and 4-Choice RT (four stimuli and two responses) of 10 karate kumite, 10 karate kata, and 10 beginners were recorded during a single session. Four karate kumite offensive actions ("stimulus") (front arm punch, front leg kick, back arm punch, and back leg kick) were projected in a real size, after which the participants performed a defensive action ("response") (high and low block for front and back attacks, respectively). The results indicated that (1) the video-based method was sensitive to discriminate between beginners and karate practitioners (p < 0.05) (higher RT for beginners), (2) no differences in RT were observed between kumite and kata (p = 0.504), (3) Simple RT was lower than 2-Choice RT and 4-Choice RT (p < 0.001) but no differences were observed between 2-Choice RT and 4-Choice RT (p = 0.112), and (4) the RT was higher for the back attacks compared to front attacks (p < 0.001). These results support the video-based method for assessing RT during karate kumite situations.