Abstract Pyrolysis is the key step in biomass thermochemical conversion process. The network model can accurately predict the pyrolysis process but generally cannot incorporate the combustion and gasification sub-models due to its complexity. This paper used the Bio-CPD model to predict the pyrolysis products of softwood and hardwood respectively; based on the predicted results, two empirical-simple forms of pyrolysis models were further optimized. The ultimate kinetic parameters obtained are suitable for biomass pyrolysis at high heating rate. For softwood, after being optimized, the apparent frequency factor and E/R of single rate are 4.3106e + 07 s−1, 10042 K respectively. While for two-competing rates model, the parameters are, α1 = 0.75, α2 = 0.89, A1 = 7992 s−1, E1/R = 7000 K, A2 = 8.3e + 09 s−1, E2/R = 14520 K, respectively. The numerical simulation of biomass pyrolysis and combustion process were performed by using CFD code Ansys Fluent. The results reveal that the release of volatile predicted by the default parameters have a delay compared with the actual process and is not appropriate for biomass simulation, while the optimized parameters in two simple models are accurate enough to simulate the biomass pyrolysis at high heating rate (103–105 K/s).