Ethanol is considered a potential biofuel for internal combustion engines. In this study, homogeneous charge compression ignition (HCCI) simulations of ethanol engine experiments were performed using stochastic reactor model (SRM). Detailed ethanol oxidation mechanism is developed by including NOx reaction in existing detailed oxidation mechanism with 57 species and 383 reactions. Detailed ethanol mechanism with NO,; used in this study contains 76 species and 495 reactions. This mechanism was reduced by direct relation graph (DRG) method, which was validated with the experimental results. Existing Lu's 40-species skeletal mechanism with NO formation were also compared with detailed and reduced mechanisms for predicting maximum cylinder pressure, maximum heat release rate and crank angle position of maximum cylinder pressure in HCCI engine. Reduced mechanism developed in this study exhibited the best resemblance with the experimental data. This reduced mechanism was also validated by measured engine cylinder pressure curves and measured ignition delays in constant volume reactors. The results showed that reduced mechanism is capable of predicting HCCI engine performance parameters with sufficient accuracy. Sensitivity analysis was conducted to determine the influential reactions in ethanol oxidation. Results also show that detailed and reduced mechanism was able to predict NO emission in good agreement with the corresponding experimental data. (C) 2016 Elsevier Ltd. All rights reserved.