I think circular dichroism is your best bet; however, for small strands of DNA it's probably not sensitive.  Something like soft-xray scattering can give you this kind of information.  The only way traditional UVVis could resolve the difference is if the circular vs. linear forms had a particular wavelength at which they absorbed/scattered strongly that changed based on these forms. 

Do you have a bacterial DNA that you think might have circular plasmids?  If not, then most genomic DNA will be sheared into ~20 kb linear pieces by most standard DNA isolation procedures and your prep is almost certainly composed of linear pieces.  If the circular DNA is supercoiled, then it would separate from linear DNA on an Ethidium Bromide CsCl gradient.  But if the circles are are not supercoiled, then they would be indistinguishable from linear DNA by all the physical methods I know of.  Electron microscopy might be helpful if there is a large portion of the DNA that is circular.

So if I have been sended bacterial DNA, will it be sheared into linerar pieces? It has been isolated by CTAB protocol.

I was wondering that because I have a DNA probe wich recognises its complementary strand, but I am not being able to detect the bacterial DNA (which contains in its sequence that "complementary strand"). I was just wondering the reason, and thought it could be because I am getting circular DNA.

Thanks Adam, I had heard that about measuring at particular wavelenghts which could indicate if the DNA is circular or linear, but I don't know if that is possible and at which wavelenghts should I measure. I would like to know that, thanks!

I have one experience with preparing bacterial DNA with a CTAB procedure and in that case there was an enormous amount of RNA contamination in the prep.  So you may wish to run a gel to see if much of what you think is DNA (based on the A260 reading) is actually RNA.  If you are doing a Southern blot, then you are probably digesting with a restriction enzyme first which would likely linearize any circular DNA present.  If you are just trying to do a dot blot hybridization, then the presence of a large excess of RNA may be preventing the proper hybridization from occurring.