The main disadvantage of Euler-Lagrangian (multi-scale) reactor models is the computational effort required for the numerical solution. This is mainly due to the large number of particles that have to be considered. In this work different solution methods for solving a biomass pyrolysis particle model are presented. The computational time that each method requires is compared, taking into account typical process parameters and time step sizes of fixed-bed and fluidized-bed reactors. It can be concluded that, if the boundary conditions of the particle model are constant, a multi-step ODEs solver - the backward differencing formulas method - is the fastest. However, when coupling a particle model to a reactor model, these boundary conditions will change over time and the particle model should provide a solution each time step of the reactor model, usually in the order of magnitude of milliseconds. In this situation the ODEs solvers require a high computational time due to it initialization every time step and the fastest method is a novel iterative, fractional-step algorithm presented in this study, where each phenomenon is solved with an implicit, semi-implicit or explicit method depending on its characteristic time. (C) 2012 Elsevier Ltd. All rights reserved.