Abstract: Crustal anisotropy plays a crucial role in the study of seismic activity and crustal dynamic evolution, which is widely constrained based on the periodic variation of moveout of the P-to-S conversion at the Moho (PmS) with respect to the back azimuth in receiver functions (RFs). Due to the uneven distribution of seismic activity, short recording period, complexity of mediums adjacent to the Moho, and presence of ambient noise, some PmS phases in RFs have no well-determined arrivals, which impedes the use of PmS to constrain crustal anisotropy. In this study, we propose a new method of jointly applying PmS and PPmS phases to measure crustal anisotropy parameters. First, perturbations of moveout caused by presence of azimuthal anisotropy are obtained by averaging those of the PmS and PPmS after subtracting the corresponding values in isotropic cases. After that, those perturbations are plotted against the back azimuth, and fitted using a cosine function. The optimal pair of anisotropy parameters corresponds to the best fitting curve. When PPmS is characterized by a higher signal-to-noise ratio than the PmS, we can only use the former to constrain crustal anisotropy. Tests using synthetic and observed seismic records with various signal-to-noise ratios confirm the applicability of the method in measuring whole crustal azimuthal anisotropy.