The Eulerian stochastic field method is applied to the solution of the modeled evolution equation for the subgrid joint probability density function (JPDF) of the reacting scalars in a large eddy simulation (LES) of a piloted methane/air diffusion flame (Sandia Flame D). A simple model for subgrid scale (SGS) stresses and fluxes and a global four-step mechanism for combustion are combined in the formulation. Test cases with varying mesh sizes and numbers of stochastic fields were completed. The differences in the results obtained with the two g-rids were very small and this indicates that the mesh resolution was sufficient. However, incorporation of the JPDF, via the stochastic field Solution method, improved the quality of predictions significantly, particularly those quantities related to reaction, such as temperature. Eight stochastic fields were shown to be enough to characterize the influence of SGS fluctuations on filtered species formation rate to reasonable accuracy and at moderate computational cost. With the exceptions of H,) and CO, good agreement between measured and computed mean and RMS profiles of velocity, composition, and temperature was achieved. The discrepancies in H-2 and CO concentrations are attributable to limitations in the global chemistry mechanism used in the LES. Overall the results serve to highlight the potential of the Eulerian stochastic field method in LES. (c) 2006 The Combustion Institute. Published by Elsevier Inc. All rights reserved.