Synthetic aperture radar Tomography (TomoSAR) provides scene reflectivity estimation of vegetation along elevation coordinates. However, the more multi-baselines acquisition, the longer the time span of acquisition, which will result in serious temporal decorrelation in forested area. In this way, we expect to use as smaller number of baselines as possible to obtain high estimation accuracy in elevation direction. We thus investigate the performance of Polarimetric SAR tomography (Pol-TomoSAR) with small number of baselines in forested areas. The results show that compressive sensingbased Pol-TomoSAR has higher estimation accuracy in elevation direction than conventional Pol-TomoSAR with small number of baselines.