In this work we perform a study of the data retention behavior of silicon nanocrystalline flash memories. Charge loss is modeled through direct and trap assisted tunneling from the nanocrystals to the channel and to the neighboring nanocrystals. The discrete loss of charge is modeled by a Monte Carlo algorithm. In addition to being more realistic, the Monte Carlo approach, can inherently take into account statistical fluctuations among different memory devices and the effect becomes more important as the devices are scaled down in size. The simulated charge retention data has been fitted to experimental data and show reasonable agreement for various temperatures and oxide thicknesses. (C) 2010 Elsevier Ltd. All rights reserved.