We examine several computational strategies for the prediction of the (17)O-NMR shielding constants for a selection of organic acids and peracids in aqueous solution. In particular, we consider water (the solvent and reference for the chemical shifts), hydrogen peroxide, acetic acid, lactic acid and peracetic acid. First of all, we demonstrate that the PBE0 density functional in combination with the 6-311+G(d,p) basis set provides an excellent compromise between computational cost and accuracy in the calculation of the shielding constants. Next, we move on to the problem of the solvent representation. Our results confirm the shortcomings of the Polarizable Continuum Model (PCM) in the description of systems susceptible to strong hydrogen bonding interactions, while at the same time they demonstrate its usefulness within a molecular-continuum approach, whereby PCM is applied to describe the solvation of the solute surrounded by some explicit solvent molecules. We examine different models of the solvation shells, sampling their configurations using both energy minimizations of finite clusters and molecular dynamics simulations of bulk systems. Hybrid molecular dynamics simulations, in which the solute is described at the PM6 semiempirical level and the solvent by the TIP3P model, prove to be a promising sampling method for medium-to-large sized systems. The roles of solvent shell size and structure are also briefly discussed.