This could be a problem with how you ran your simulation, but it could equally be a problem with your calculation/interpretation of diffusion coefficient. You'll need to provide more info. Consider these while you provide more info.
(1) Are you sure you are getting enough sampling? How is the ergodicity of this system? Perhaps look at (a) energy vs. simulation time, (b) your MSD plots, (c) the velocity autocorrelation function
(2) How is the the diffusion coefficient that you "want" acquired? Is it from simulation or experiment? If its from simulation, make sure every setting is the same and try to match it before you start customizing your input. If its from experiment, then a difference of that order is very understandable. Experimental diffusion coefficients can be very inaccurate depending on the nature of the system or the type of measurement, and even when the experiment is very carefully performed, it can still be different from theory.
(3) What sort of statistically sampling are you using to extract the diffusion coefficient?