As one of the most basic parameters, manikin body feature could be an important factor influencing the airflow and temperature fields in indoor environments. This study aims to improve the computational efficiency by optimising and simplifying manikin body features. A 3D scanned computer-simulated person (CSP) with extremely detailed body features was employed, followed by two simplified CSP models with different approaches. One of the simplified models was rebuilt based on the skeleton of the 3D scanned model with very limited body features, while the other model was simplified by removing some of the features from the 3D scanned model. All CSPs were tested under quiescent condition, followed by further comparisons under displacement and mixed ventilations. The outcomes indicated that the geometric difference of manikin body would have significant impact on the airflow patts near manikin bodies, whilst it has very limited influence on the temperature field. The difference of body features could significantly affect the development of thermal plume, which mainly reflected above the manikin head. Also, change of CSP body features due to simplifications may become more sensitive to the predicted results under mixed ventilation, as a result of fewer interactions between the thermal plume and injected airflow.