Fuel stratification introduced by direct injection (DI) of isooctane in an optically accessible stratified charge compression ignition (SCCI) engine is numerically investigated using a multidimensional model. The study is motivated by the fact that in homogeneous charge compression ignition (HCCI) engines operated at low load, combustion efficiency is rather low and the associated carbon monoxide (CO) and unburned hydrocarbons (UHC) emissions are quite high but that this can be improved by fuel stratification using DI. The resulting in-cylinder mixture distribution is crucial to the success of this strategy. Regions that are too rich cause high NO infinity while fuel in regions that are too lean results in deteriorated combustion efficiency and OC/UHC emissions. Methods to predict the fuel distribution are therefore required. This study aims to determine the extent to which a computational fluid dynamics (CFD) model can predict fuel stratification in SCCI engines and determine whether the predicted in-cylinder fuel and temperature distributions can explain emissions trends with different stratification levels. The model is shown to have quantitatively good agreement with experimental measurements of the fuel distributions for various injection timings under nonfiring conditions this is apparently among the first such demonstrations in SCCI operating conditions. It is found that with more retarded injection timing, fuel is increasingly concentrated in the central regions, leading to potential improvements of combustion efficiency and reduction of CO and UHC. However, nitrogen oxides can be potentially increased due to the appearance of regions with excessively high equivalence ratios. The creation of high equivalence ratio regions was examined and it was found that spray-to-spray interaction and spray-wall interaction play important roles in mixture formation. The sensitivity to model parameters was also examined.