<p>The effect of forest structure affected productivity of harvester operator work in a continuous cover forest. In particular, the effects of the forests consisting of even- and uneven-age trees were compared. In addition to tree variables describing the structure of the forest, variables describing the time taken in each of the work phases were needed to accurately model the logging motions. Modelling also required single tree data from the harvester. The model precisely predicted the relationship between the variables and productivity in both the uneven-aged forest (<i>R</i>^<i>2</i> = 0.96) and the even-aged forest (<i>R</i>^<i>2</i> = 0.92). Both explanatory powers of models were statistically significant. Productivity was explained by the “volume of trunk”, the “length of the operating part of trunk”, the “moving time of the logging device to trunk”, the “felling time of tree” and the “processing time of trunk”. In the uneven-aged forest, the effective-hour productivity was 37.7&#xa0;m³&#xa0;<i>E</i>_<i>0</i><i>h</i>⁻¹ and in the even-aged forest 43.5&#xa0;m³&#xa0;<i>E</i>_<i>0</i><i>h</i>⁻¹. The work phases “moving the logging device to trunk” and “felling of tree” consumed more time in the uneven-aged forest. The results of the time and motion analysis justify the promotion of training both work phases to increase productivity of harvester operator work. This modelling approach can be recommended for studies on the development of selective logging method for continuous cover forestry.</p> Graphical Abstract <p></p>