The difficulties in quantitation of in vivo 31P spectra are exacerbated by the fact that, in general, coils with inhomogeneous B1 fields are used with in vivo samples. A general method for quantitation of in vivo 31P MRS results obtained with the ISIS localization method was developed using computer simulations. The simulation calculates the preparation of the sample magnetization throughout the sample by the ISIS pulse sequence, as well as the sensitivity of signal reception. The calculation accounts for both the B1 field and the B0 gradients applied to the sample. The sensitivity of the experiment is expressed by integration of the simulated signal over the sample, assuming a homogeneous sample. The primary advantage of this approach is that a separate localization experiment on a phantom of known concentration is not required each time parameters of the localization experiment, such as dimensions or location of the localized volume, are altered. In addition, the simulations indicate the degree of contamination (signal from outside of the localized volume) that occurs, and provide a means of comparing different executions of the ISIS experiment. Experiments were performed on phantoms to verify the simulations, and experimental results on human brain and liver are reproduced to show that this approach provides reasonable estimates of metabolite levels in terms of molar concentrations.