In this study, we assess the quality of the digital surface model (DSM) generated from Pleiades-1 tri-stereo images which can potentially contribute to the detection of citrus trees in terms of height information. The methods tested on stereo/tri-stereo images are; (a) local methods (correlation-based and least squares method), (b) semi-global method (semi-global matching (SGM)), and (c) global method (SIFT-flow). DSMs of three sub-regions of Mersin area have been presented for each stereo/tri-stereo matching method; whereas for the SIFT-flow method, we have only depicted the parallax results. Numerical results reveal that the SGM forward-backward stereo combination which has the largest intersection angle provided the best results in 2 out of 3 test areas. However, the results confirm that none of the methods could reach the desired level of performance for the height estimation of citrus trees that can potentially guide the detection step.