Thermal stability is an important and interesting physical property of proteins. A common method to study it by is circular dichroism (CD) spectroscopy. The aim of this study was to test methods to improve thermal stability analysis by CD spectroscopy. Experiments were performed using the Abp1p SH3 domain from yeast as a model protein. Thermal denaturation was monitored at multiple wavelengths. It was concluded that for data sets of reasonable quality the choice of wavelength does not affect the results. An approach to estimate stability of thermophilic proteins was tested where thermal stability was measured at different concentrations of the denaturant GuHCl. The thermochemical data was used to estimate the stability in absence of GuHCl by extrapolation. The results were compared to those obtained from CD spectroscopy and differential scanning calorimetry. It was found that a stabilizing effect from low concentrations of GuHCl complicated the extrapolation. It is likely that this method is more successful if there is no stabilizing effect. The effect of ΔCp in stability parameter calculations was investigated with an experimentally and theoretically determined ΔCp. This was further investigated with synthetic data sets. The ΔCp used in calculations had no notable effect, as long as there was no cold denaturation. Although ΔCp is not necessary in calculations, it is an interesting parameter itself. ΔCp can be calculated from the thermochemical data used for extrapolation. The results in this study demonstrate robustness in thermal stability analysis by CD spectroscopy and a potential for development.