Uncertainty analysis of laminar and turbulent aeroheating predictions of a 70 degrees spherically blunted cone at peak heating condition during Mars entry is performed. Due to the difficulty in obtaining accurate chemical kinetic model, numerical aeroheating predictions possess significant uncertainties. In this study, 14 rates in Park 8-species 14-reaction model are treated as epistemic uncertain variables represented with intervals, meanwhile stochastic expansion based on point collocation non-intrusive polynomial chaos expansion, is utilized to represent and propagate the uncertainties. In particular, Sobol indices are used to rank the relative contribution of each rate. 240 CFD evaluations are employed to obtain the laminar and turbulent uncertainty results respectively. The results show that the top contributing parameters to turbulent heat flux are similar to those observed in the laminar case in the windside region. However, in the leeside region, the key parameters, which produce significant uncertainties in laminar and turbulent cases, are evidently different. The maximum uncertainty in laminar aeroheating prediction is about 6%. In the large region of leeside flank, the uncertainty in turbulent aeroheating prediction is above 7%. Moreover, the uncertainty interval is as wide as 32.4 W/cm(2), which is about 15% of the mean value. (C) 2017 Elsevier Ltd. All rights reserved.