Generally most of the visualization tools able to open Gaussian log files (Gaussview, Molden or Avogadro) automatically recognize the type of bonds from their lengths; while this is pretty easy and affordable for simple organic molecules this kind of automatic identification may lead to errors in more complex cases. Of course this depends on the particular algorithm used for the identification of the bonds. This is only for a rough visualization and it has nothing to do with theoretical approaches for the definition of bonds between atoms.   

Dear Smruti,

In order to see the bond order you should perform a NBO analysis. There, if you search for the orbital description of each bond, you will see that sometimes there are two or three bonds that are numbered with its portions of the orbitals from each atom.  (For example, in a double bond you should expect the sigma bond with a covalent polarization of the p orbital and below of this, the PI bond with higher polarized p bond).

I hope it helps you,

Joaquim Rius

Thank you Joaquim.

Thanks Joseph.

You can also print 'bond-order' information. For instance, Mayer bond-orders are available using IOp(6/80)=1 with Gaussian. For analysis check slide #96 here:

http://stark.udg.edu/~eduard/quimio1.pdf

Dear,

The only problem I see in only using the Gaussview is that the "automatic" bond figure depends on the distance. Maybe the approximation is true for a very covalent compounds or carbon compounds but for complex structures it is very different.

For example, in a coordination compound. I has a case where the bond represented is single but the NBO analysis revels a triple bond! In order to understand properly a bond I think that it is a requirement to compute some kind of population analysis because, on the other hand, with the general or descriptive  knowledge in chemistry is more than enough and useful and with more critic point of view. In fact, with Lewis structures, resonance structures and a piece of paper you can get more valuable information than from the only Gaussview "representation".

 Agree with Joaquim. If you want a 'Lewis picture' of your molecule I would go for a Mayer bond order analysis, which is also much simpler than a NBO analysis. Only be aware: if you have diffuse functions in your basis set you might need to perform a 3D atomic partition because Hilbert-space population (and bond-order) analysis are not reliable with diffuse functions. The latter can be performed with the wfn output from Gaussian using the programs of Prof. Mayer: 

http://occam.chemres.hu/programs/

 Yes, from the log files, it is very easy to see the distance between two bonds. But, is there any particular value( or defined value) for single or double or others for a sample molecule?

Please explain Amy J. Austin.