Lignocellulosic biomass (LB), such as agricultural residues, is the first source of biobased carbon on Earth. Made up mainly of plant cell walls, LB can be valorised into biobased products thanks to transformation processes combining physical or physico-chemical pre-treatments and hydrolysis of the polymers by enzymes. This valorisation is a promising way of reducing our dependence on the petrochemical industry but is still limited by the low efficiency of the enzymatic hydrolysis step due to the natural recalcitrance of LB. This recalcitrance is mostly induced by the tight association of polysaccharides (cellulose, hemicelluloses) with lignin and hydroxycinnamic acids within the cell walls, which prevents the cell wall deconstruction by limiting the polysaccharide accessibility to hydrolytic enzymes. Recalcitrance varies according to the plant species, to the tissues but also to the cell type. The hot water pretreatment (HWP) is an eco-friendly process, which reduces the hemicellulose content and increases the enzymatic conversion capacity of LB. Its impact is the subject of numerous studies but the physico-chemical modifications induced on different properties of LB and the links with enzymatic hydrolysis are still unclear. The objective of this work is to combine global biochemical and physico-chemical analyses (composition and 13C CP/MAS NMR) with cellular scale analyses using imaging and microscopy approaches to study the effect of HWP according to cell variability. Data show that the HWP induces a reorganization of the lignocellulosic network through: i) modification of the cell wall composition with a loss of hemicelluloses, hydroxycinnamic acids; ii) modification of polymers conformation highlighted by an increase of the apparent crystallinity and a decrease of the ether bounds on S and G lignin units. These modifications have allowed to improve the mobility of fluorescence probes, measured by FRAP, translating looseness within the lignocellulosic matrix and more accessibility to the cellulose, which promotes enzymatic hydrolysis. Imaging and microspectroscopy approaches clearly highlight the different behaviours of different cell types with regard to enzymatic hydrolysis as well as HWP. We also made clear that the parameters reported as major contributors to recalcitrance (lignin content, crystallinity, etc.) do not necessarily have a negative impact on hydrolysis capacity when their value is increased. Overall, we highlight that in order to better understand the factors linked to recalcitrance and the modifications following HWP, it is necessary to take into account the complexity of the lignocellulosic biomass through multi-scale analyses.