The purpose of this study was to analyse the fundamental phenomena related to the swelling behaviour of densified and thermally modified wood under changing moisture content, as well as to investigate the differences in recovery behaviour between four different methods: (i) soaking-drying cycles, (ii) soaking-drying cycles in hot water, (iii) water-soaking with continuous data logging to measure dynamic changes in thickness and (iv) humid-dry cycles at different relative humidity (RH). Methods were applied to untreated, thermally modified, densified (D) and densified + thermally modified (DTM) Scots pine sap wood samples. Soaking methods were found to produce significantly higher set-recovery results than RH cycling, with higher temperature accelerating the relaxation during soaking. Repeated cycles increased the swelling of the densified samples. The RH thresholds for set-recovery were found to be between 65 and 75 % for the D samples and between 75 and 84 % in the case of DTM samples.