A thermally thick particle model is proposed and combined with the Euler-Lagrange model to study the biomass pyrolysis process in a fluidized bed reactor. Besides model validation, both a single biomass particle and a batch of particles are simulated. The large differences between thermally thick and thermally thin models are also highlighted by several indicators: particle surface and core temperature evolutions, mass loss history, particle trajectory, and product gas distributions. Results show the importance and necessity of thermally thick treatment for large particles because there exist big intraparticle temperature gradients, which in turn make different parts of the particle experience different conversion stages. Such behavior cannot be predicted by a thermally thin model. In addition, effects of particle size and operating temperature are also explored, revealing that smaller particle size and higher temperature promote the pyrolysis process and reduce the time period during which the core temperature approaches the surface temperature.