Anthracology has been traditionally focused on the botanical identification of the charcoal fragments and, therefore, the observation of a generalized pattern in the use of wood species in different combustion structures has been understood as the absence of selection criteria. However, ethnographic studies reveal the existence of an organized firewood gathering activity oriented to the different physiological and phenological states of the wood: green, drift, healthy, dead and rotten wood [1,4]. In order to characterize the macroscopic state of the firewood used by hunter-gatherer groups during the past, several experimental and ethnoarchaeological studies have been carried out shedding light on the palaeoeconomical approach from charcoal analysis [3,5]. We present here the results of the charcoal analysis carried out in four Middle Palaeolithic combustion structures from two sites located in Eastern Iberia: combustion structures H4 and H11 from Abric del Pastor, unit IV (> 75 ka BP) and combustion structures H50 and H57 from El Salt, unit Xb (ca. 52 ka BP). In Abric del Pastor Juniperus sp. (junipers) and Pistacia sp. (terebinth) were the main taxa while in El Salt most of the firewood used was from Pinus nigra-sylvestris (Black/Scots pine) and Acer sp. (maple). We defined a protocol for the microscopic observation of the anatomical alterations caused by fungi activity on wood before combustion processes. We classified the wood charcoal fragments in three alteration levels (A.L.) and we calculated the alteration index (A.I.) according to the previous experimental studies . The classification of wood charcoals according to their alteration degree revealed differences between the hearths studied: while combustion structures H4, H11 and H50 showed values equal or lower than 0.35, hearth H57 showed a considerably higher degree of alteration (0.46). The comparison between archaeological remains and the previous experimental and ethnographic data point out to several firewood acquisition criteria: the low values of alteration from the first three combustion structures suggest a firewood gathering activity based on the collection of dead wood branches regardless of the taxa. On the other hand, hearth H57 showed an Ai which corresponds to the limit between the dead and rot wood of the experimental study. In addition to this, a total of 253 seed fragments of cf. Acer sp. were identified inside the structure suggesting the hypothesis of the use of green Acer sp. wood with the seeds still inserted in the cut branches. In the smoking hearths from ethnographic contexts the humidity rate of the wood is more important than the species itself and large part of the fuel is composed of green wood with the addition of dead wood that makes combustion to function . Thus, the selection of green wood mixed with very altered or rotting conifer firewood could be related with a function of the structure to produce smoke. Finally, the production cycle of the maple, beside the nut-shells preservation inside the combustion structure, has given us some clues about the seasonality pattern during the occupation event related to hearth H57. In conclusion, the applied methodology has allowed us to detect nuances between the anthracological record of the different combustion structures that would not be visible by merely following the traditional anthracological methodology. Furthermore, these data support the idea, experimentally and ethnographically defined, that there is no particular ”good fuel” but a diverse range of physiological and phenological states of the wood linked to the aims pursued and the fuel needs.