Steel is one of the most important materials used in modern society. The majority of the steel produced today is based on the use of coke and contributes a lot to greenhouse gases emission. Many researchers have been laid on the possibility to replace part of the fossil-based energy source in iron making with renewable, biomass-derived reducing agent. The main problems of this replacement are some difference of in quality between coke and wood charcoal (more reactive, less strength and carbon content) It causes a little shutdown of production in blast furnace and additional cost to modify a furnace. The aim of this paper was to determine in a statistical manner how carbonizations parameters impact the charcoal quality, especially reactivity and mechanical parameter. We applied a random factorial design and used the General linear System procedure to perform the statistical analysis. The experimental study was carried out using Eucalyptus Urophylla and Eucalyptus Camadulensis wood and involved two carbonization temperature (350 and 600°C), two relative working pressure (2 and 6bars) and two heating rates(1 and 5°C/min). Six response variables were analyzed and discussed following a random factorial design: the charcoal yield (y char), the fixed carbon content (Cf), the bulk density (D), the compressive strength (Rm), friability (F) and the reactivity(R) of charcoal. Except for the friability of charcoal, all other property are well correlate with carbonization parameter. In the range of low carbonisation parameter, reactivity of charcoal is affected only by carbonization temperature.