The pyrolysis kinetics of beech wood was analyzed using model-free and model-fitting methods. Experimental measurements of the pyrolysis process were conducted in two thermogravimetric analyzers (TGA), a TG 209/2/F from Netzsch and a TGA Q500 from TA Instruments, which were found to have a similar precision in the establishment of the present heating rate. Two experimental procedures were employed: (i) introducing samples which were pre-dried externally before the experiments were executed and (ii) internal (in situ) drying of the samples in the TGA via a special temperature program below 150 degrees C which preceded the pyrolysis process. The kinetic parameters were derived (i) using several model-free methods, namely Kissinger method, iso-conversional methods, a simplified Distributed Activation Energy Model (sDAEM) and, (ii) using a model-fitting method via a five-step reaction model which calculates the differential thermogravimetric (DTG) curves at different heating rates; the calculated DTG curves were further analyzed by Kissinger's method to obtain overall kinetic data. The kinetic parameters were found to be different in the two experimental procedures. Also, they turned out different when the assumed end temperature of the pyrolysis process was varied. This is because the pyrolysis of slowly charring solid residues becomes more important with increasing temperature and finally overruns the release of volatiles from the wood samples. For the same experimental procedure and for sufficiently low end temperatures, corresponding to a degree of conversion less than 85%, model-free and model-fitting methods resulted in similar kinetic parameters.