The ability of conventional electron correlation (MP2 and QCISD) and density functional theory (B3LYP and B3P86) methods to provide accurate and reliable optimized structures, and homolytic S-N bond dissociation energies (BDEs), for a range of S-nitrosothiols (RSNOs) has been investigated. It is found that, in general, for any given method the 6-311 +G(2df,p) or larger basis set must be used to obtain reliable structures. With a suitably large basis set, the different methods generally give optimized structures in close agreement with each other. However, the B3LYP method consistently overestimates the RS-NO bond length. The trends observed are found to be due in part to the fact that the RS-NO bond does not possess considerable double-bond character as previously suggested, but rather is a long single S-N bond, with the -NO moiety possessing considerable multiple-bond character. The B3P86/6-311+G(2df,p) method consistently gives BDEs in best agreement with values obtained with higher accuracy methods, e.g., CBS-Q, while the B3LYP method increasingly underestimates BDEs with increasing RSNO size. In contrast, for all RSNOs, the QCISD method significantly underestimates BDEs by as much as 55 kJ mol(-1). Overall, the B3P86/6-311+G(2df,p) method is found to perform the best of the methods considered for obtaining optimized structures and homolytic S-N BDEs of S-nitrosothiols.