Emission lines from metals offer one of the most promising ways to detect the elusive warm-hot intergalactic medium (WHIM; 105 < T < 107K), which is thought to contain a substantial fraction of the baryons in the low-redshift Universe. We present predictions for the soft X-ray line emission from the WHIM using a subset of cosmological simulations from the Overwhelmingly Large Simulations (OWLS) project. We use the OWLS models to test the dependence of the predicted emission on a range of physical prescriptions, such as cosmology, gas cooling and feedback from star formation and accreting black holes. Provided that metal-line cooling is taken into account, the models give surprisingly similar results, indicating that the predictions are robust. Soft X-ray lines trace the hotter part of the WHIM (T > rsim 106 K). We find that the OVIII 18.97 Å is the strongest emission line, with a predicted maximum surface brightness of ~102photon s-1 cm-2 sr-1, but a number of other lines are only slightly weaker. All lines show a strong correlation between the intensity of the observed flux and the density and metallicity of the gas responsible for the emission. On the other hand, the potentially detectable emission consistently corresponds to the temperature at which the emissivity of the electronic transition peaks. The emission traces neither the baryonic nor the metal mass. In particular, the emission that is potentially detectable with proposed missions traces overdense (ρ > rsim 102ρmean) and metal-rich (Z > rsim 10-1 Zsolar) gas in and around galaxies and groups. While soft X-ray line emission is therefore not a promising route to close the baryon budget, it does offer the exciting possibility to image the gas accreting on to and flowing out of galaxies.