Traditional smoke visualisation is based on stationary objects with a moving fluid. This paper presents wake flow from moving manikins investigated by a new smoke visualisation technique. Three scaled models 1/5th of realistic manikin models were used, including two body shapes (thin and wide) and two gestures (standing and walking). Smoke visualisation was produced by a chemical reaction between acetic acid (CH3COOH) and cyclohexylamine (C6H13 N) to generate smoke. Highspeed photography and image processing techniques were used to determine the qualitative and quantitative data on the airflow patterns and separation points from the manikin motion. Detailed flow separation images showed that regular vortices were produced off the head and the shoulder, while flow separating at the hand swirled behind and inwards. Analysis of flow over the head revealed how the separation point shifted from the back of the head to the front as the velocity decreased. When the manikin came to a stop, the results also showed increased airflow activity around the face which increases the risk of inhaled particulate matter from the surroundings. Image processing analysis produced quantitative data which included the vortex shedding frequency coming off the shoulder, and the separation point from the manikin head. It is expected that the experimental results will provide validation data for future computational fluid dynamic (CFD) modelling results.