For the ten last years, many laboratory studies have showed that coke reactivity and its strength after gasification are mainly dependant of the coal rank. We confirmed these results for semi-industrial cokes, and explained them by the measure of the coke microtexture (ref. 1). It was found not possible to predict reactivity only with coal rank. Other parameters as plastic properties or chemical analysis of ashes have to be discerningly used to predict coke reactivity. In the present paper we show that to apply these laboratory scale theories to the industrial production, it is also necessary to take into account carbonization conditions (chamber width, heating wall temperature and heterogeneity in the oven).