Terrestrial laser scanning is an accurate and relatively fast method for point clouds acquisition, which are often used for 3D forest modelling and estimation of selected dendrometric parameters in forestry. In the dissertation thesis, we focused on aspects of point cloud processing, which are important for estimation of longitudinal shape of the trunk of standing trees. We have solved the problem of optimizing the point cloud horizontal cross section width for estimation of tree diameters and accuracy of diameter in different heights above the ground estimation. From the acquired data, we calculated the stem profiles, shape quotients, false shape series, convergence of trunks and slenderness coefficients of trees.Data collection took place on a circular research area with a radius of 16 m. Sessile oak (Quercus petraea (Matt.) Liebl.) with an average height of 29,6 m and an average diameter at breast height of 31.8 cm was 100% represented in the research area. We scanned the research area with the multi-scan method using a Faro Focus3D 120 terrestrial laser scanner from seven positions and with Trimble TX8 terrestrial laser scanner from thirteen positions. Programs recommended by scanner manufacturers were used to reference the scans and initially filter the point clouds. The point clouds were then processed in DendroCloud 1.50. Unsuitable points were filtered from the point clouds, a digital terrain model derived, cross-sections 1.3 meters above the ground created, and tree trunks identified. In the next steps, 19 cross-sections of various widths, from 1 cm to 100 cm, at a height of 1.3 m, were created from the point cloud generated by the FARO Focus3D 120. In the case of a point cloud created with a Trimble TX8, the longitudinal shape of the trunks was characterized by the creation of 16 sections with a cross-section thickness of 10 cm at heights from 0.3 m to 13 m.Point cloud cross-sections with different widths showed a systematic underestimation of the values measured by diameter tape by an average of 1.05 cm and a root mean square error of estimating the tree diameters from 1.04 cm to 1.23 cm. Tree diameters estimated from a point cloud at different heights above the ground were compared with reference values measured non-destructively using lifting platform and diameter tape. Statistical evaluation of the data confirmed the significant effect of the height of point cloud cross-section on the estimation of tree diameter. At a height of 0.3 m, the irregular shape of the tree trunks had a significant effect on the derivation of the diameter. At all point cloud cross-section heights, except 0.3 m, the derived values underestimated the measured values by an average of 0.62 cm. The root mean square error of estimating the diameter of the trees depending on the height of the cross-section varied from 0.8 cm to 2.21 cm. From the reference and estimated tree diameters at different heights, the characteristics of the longitudinal shape of the trunk were subsequently constructed: trunk profiles, convergence, slenderness coefficient, shape quotients and false shape series.The achieved results confirm the applicability of the multi-scan method of terrestrial laser scanning in forestry, not only in estimating the commonly used diameter of trees at a height 1.3 m above the ground, but also in estimating the longitudinal shape of tree trunks. Estimation of the longitudinal shape of a standing tree trunk from point clouds is a promising method to refine the methods used in forest inventory and the creation of detailed 3D models of standing trees.