Thermal modification is an efficient way to improve some wood properties, like equilibrium moisture content (EMC), dimensional stability and durability (Esteves and Pereira 2009), to produce a wood modified material. According to the literature (Candelier et al. 2016), it has been observed that the thermal degradation of wood has a high dependence on the initial wood characteristics (wood specie, density) and process parameters, such as drying stage, heating medium, and treatment intensity (heating rate, temperature and duration). The objective of this study is to predict the treatment duration in order to reach a particular level of wood modification under industrial conditions. For that, the mass loss dynamics during the treatment are recorded and modelled. The flow chart of research methodology is shown in Fig. 1. The obtained results are encouraging for a future development of a numerical tool able to conduct performing the heat treatment of wood in industry. The experiments were carried out on wood boards (30 x 22 x 2.5 cm3) in a pilot scale system, under conditions close to the industrial ones (heat transfer by conduction with 0.2°C min-1 heating rate under vacuum). Two different wood species, a hardwood: the poplar (Populus nigra) and a softwood: the fir (Picea abies), were used. The heat treatment was conducted at 230 °C for 14 hours under low air pressure (200 hPa) with less than 5% oxygen content. 14 hours duration is examined to obtain kinetic profiles required for the modelling, while usual treatment duration in the industry is close to 1 to 5 hours. Results of the present work are shown in Fig. 2. They indicate that the mass loss of poplar (14.21 wt%) is higher than fir (10.45 wt%). The difference of thermal sensitivity between poplar and fir is due to the hemicelluloses composition of hardwood and softwood (Chaouch et al. 2010). Moreover, if the target of the wood modification is to reach 10 wt% of mass loss, the duration for the poplar is closed to 75