The Planetary boundary layer (PBL) plays an important role in energy exchange and pollutant dispersion. The height of the PBL (PBLH) usually exhibits distinct diurnal variation. In this study, we investigate the char- acteristics of PBLH variation of Beijing China based on 1-year observations from Micropulse LIDAR (MPL) data, and compared against the PBLH simulated by the Weather Research and Forecast (WRF) model using eight different schemes. The variation of daytime MPL PBLH has been classified into four representative types using k- means clustering technique. Through comparison against MPL, it is shown that the Mellor-Yamada Nakanishi and Niino Level 3 (MYNN3) scheme, a second-order closure scheme, performs the best in representing the PBLH variation in most of the cases. The non-local schemes also show advantages in convective situations, due to the stronger vertical mixing embedded in these schemes. Further analyses show that differences in PBLH definition is a major factor causing the divergence among different schemes, while the biases of the profiles play a minor role.