Using the potential surface information for (KCI)(5) and Ar-9 and partition function models introduced in the preceding paper [Ball and Berry, J. Chem. Phys. 109, 8541 (1998)] we construct a stochastic master equation for each system using Rice-Ramsperger-Kassel-Marcus (RRKM) theory for transition rates between adjacent minima. We test several model approximations to reactant and transition-state partition functions by comparing their master equation predictions of isothermal relaxation for (KCI)(5) and Ar-9 with the results of molecular dynamics simulations of relaxations performed in the canonical ensemble. Accurate modeling of the transition-state partition functions is more important for (KCl)(5) than for Ar-9 in reproducing the relaxation observed in simulation. For both systems, several models yield qualitative agreement with simulation over a large temperature range. This full treatment of small systems using realistic partition function models is a necessary first step in the application of the master equation method to larger systems, for which one can only expect to have statistical samples of the potential energy surfaces.