First, treat the unreduced protein with N-ethylmaleimide (NEM) to block any cysteine that is not in a disulfide bond. Remove excess NEM (you can use a gel filtration spin column). Add TCEP reducing agent to the protein to reduce the disulfide bond if there is one. React the protein with fluorescein maleimide. Run the protein on SDS-PAGE. If there was a disulfide bond, the protein will be fluorescent. As a negative control, omit the TCEP reduction.
Running SDS PAGE with reduced and non reduced samples in sometimes successfull : it depends on the impact of the presence of the disulphide bridges on the overall architecture of the denatured protein... If there is an impact, the reduced form of the protein will run faster than the non reduced one.
Adam outlined a very nice method for visualizing the presence of disulfides, although I would add that if your disulfides are intramolecular, TCEP might not be able to reduce these due to its bulky nature, ergo, it would be more reasonable to carry out the reduction using mercaptans (like DTT).
In our lab, we routinely use mass spectrometry to check for disulfides. The principle is very simple: One sample is reduced with DTT and another one remains unreduced. Then we run these through a C4 column in acidic conditions using an UPLC system and measure the masses (we use ESI-TOF MS for this). From the chromatogram you will already be able to see if something happens in terms of equilibrium or sample homogeneity, although I don't trust this quantitatively like I trust gel filtration. From the masses you get, you will easily see if your protein is reduced from a dimer to a monomer, for instance, or if a monomer with an intramolecular disulfide is reduced (mass increases by 2 Da, due to the two extra hydrogens when -S-S- becomes 2 x -SH). Mass spec is great because you need very little sample and when using an LC step the sample doesn't need to be exactly pure. One should also remember that especially intermolecular disulfides are rather common artifacts arising in recombinant protein preparations without reducing agents.
DTNB is not a very sensitive method for detection of thiols, requiring several micromolar., which might be a limitation for a protein. Thio-Glo is a fluorogenic thiol detection reagent that is more sensitive.