Data-driven agriculture depends on management based on environmental information such as ambient temperature, humidity, solar radiation, and others. The inclusion of plant information such as photosynthetic ability, transpiration rate, respiration activity, and metabolic and physiological responses would improve quality and yield and facilitate disease controls. This research aimed at detecting plant physiological responses with the use of laser speckle analysis, which can detect microscopic motion that is invisible to the naked eye and might detect motion due to physiological responses. We investigated its ability to detect plant water stress, and how best to analyze speckle images of soybean’s leaf. We compared the Fujii method, which is fast and accurate at low brightness, and the Generalized difference method, which can distinguish long temporal fluctuations clearly. Both methods could detect the physiological response to water stress as decreasing bioactivity, however the Generalized difference method could detect faster than Fujii method. Because the detection accuracy depended on the analysis method, it will be necessary to select the method according to the target physiological response.