If you already have the trajectory the electrical conductivity should be calculated directly by tracing the ionic displacement. See equation 2 here:

http://arxiv.org/pdf/1302.4563.pdf

Otherwise you should do a non-linear MD. To suggest you a tool I need to know what is the format of the trajectory or what MD software did you use.

Dear Vesselin,

I am using DL_poly NVT simulation.  I usualy use tools like travis analyzer or vmd plugging to do my analysis. 

I guess I should go into scripting for this.

I've just spoken to colleagues who are used to use DL_POLY but I didn't got any positive response. So maybe you should do some scripting. Another way (people say that it might help) is to use the estimated D:

http://www.sciencedirect.com/science/article/pii/0167273895001723

If your system is homogeneous and the diffusion coefficient don't suffer over or underestimation as it described here:

http://pubs.acs.org/doi/abs/10.1021/jp0477147

I think the results derived by tracing ions and direct implementation of Nernst-Einstein formula will be very similar.

I was planing to get the molar conductivity with the Nernst-Einstein equation also as my final aim is to estimate the ionicity of my system.

Thanks a lot