This paper introduces a 'super drop' concept for the characterization of a sprinkler spray, where the statistical characteristics of each 'super drop' have been determined from the experimental data of a Wormald 'A' cu/p sprinkler in the pendant position. The computational fluid dynamics (CFD) model JASMINE has been adapted to incorporate a particle tracking formulation based on the Particle-Source-In-Cell method and a sprinkler sub-model based on the 'super drop' concept. The enhanced JASMINE sprinkler model has been verified against a 111 scale experiment examining the influence of the water spray on hot combustion products due to a fire in an enclosure. The predicted gas temperatures and the velocities are shown to be in qualitative agreement with the measurements. Further improvements in the model predictions can be obtained by improving the heat loss treatment to the boundaries and suitable optimization of the 'super drops'.