Abstract High-resolution water sorption isotherms were measured on 13 representative mudrock samples in order to assess the mechanisms of water vapour sorption and their relationship to the pore structure of mudrocks. The isotherm measurements were performed at 303 K on a gravimetric, dynamic vapour sorption device. Experimental data were interpreted by traditional physisorption models for which the validity was evaluated by relating model parameters to those obtained from nitrogen physisorption measurements. No direct relationships with the pore structure were observed, except for the Gurvich total pore volumes and the corresponding porosity data. Specific surface areas from Brunauer–Emmett–Teller theory are ambiguous and do not relate to nitrogen data, suggesting that water molecules do not adsorb as (multi-) layers covering pore walls. The volume filling theory (Dubinin–Astakhov equation) fits the water sorption data well but no relationship to the nitrogen data was observed in the studied sample set. A lower affinity of water for micropores was evident from the higher filling pressures of N 2 -based micropore volumes. The Barret–Joyner–Hallenda theory combined with N 2 physisorption measurements on moist mudrocks revealed that capillary condensation prevails close to saturation but not below about 0.94 relative pressure ( p / p 0 ). A distinct low-pressure hysteresis was observed from hysteresis scanning that was attributed to surface chemistry since capillary condensation occurs only at very high relative pressures. Analysis of mineralogical composition, total organic content (TOC) and organic maturity in relation to water sorption revealed only a weak correlation with the total clay content. In contrast, cation-exchange capacity (CEC) strongly correlates with water uptake, which evidences a surface-chemistry-controlled sorption mechanism. Tests of the influence of the exchangeable cation were inconclusive because pore system alteration due to cation-exchange probably superimposed the effect. To further assess the sorption mechanisms of water, nitrogen physisorption isotherms were measured on moisture-equilibrated mudrocks (11, 52, 75, 94% relative humidity at 298 K). Micropore analysis and cumulative pore-size distributions denote that water blocks pore throats rather than fills pore volumes at lower relative humidities. Over the entire humidity range, no direct relationship between water sorption and pore size was observed. These findings imply that water adsorption does not sequentially fill pores with increasing radii in mudrocks as relative humidity increases, as would be expected from water sorption by capillary condensation. This conclusion has important implications for the interpretation and measurement of geomechanical and petrophysical properties of mudrocks. Capillary pressures, particularly at low water saturations, are often calculated from water saturation using a concept based on the Kelvin equation for capillary condensation. Since water sorption in mudrocks seems to be controlled by surface chemistry rather than pore size, this approach is questionable. The observations reported here suggest that the water distribution in mudrock pore systems resulting from vapour equilibration differs from that obtained by fluid displacement (i.e. capillary drainage or imbibitions). A further consequence is that water vapour equilibration is a convenient, but not necessarily representative, method to obtain partially water-saturated mudrock samples for laboratory measurement of saturation-dependent geomechanical or petrophysical properties.