Understanding the observation abilities and advantages of ceilometer and cloud radar will be helpful for their operational applications in cloud observation. During the summer of 2014, the third Tibetan Plateau Atmospheric Science Experiment was carried out and clouds were observed using Ka-band solid-state transmitter-based millimeter wave cloud radar, Lidar and ceilometer. Based on the cloud measurements in this experiment, the algorithm to determine cloud base using cloud radar observations was developed. The biases of cloud base and data acquisition ratios by ceilometer and cloud radar for low clouds, medium clouds and altocumulus were analyzed. Reasons for the overestimation of cloud base by ceilometer were explained. The improved observation mode was introduced and the observation skill was simulated. The results indicate that (1) the cloud radar observation ability for altocumulus are better than that of ceilometer, the cloud radar echoes below the ceilometer-derived cloud bases are possibly clutter, and its observation for low clouds should be improved. Blockages of low clouds affect the observations of medium cloud and altocumulus by ceilometer. Compared with ceilometer, the cloud radar missed some shallow altocumulus. (2) Most of cloud bases observed by ceilometer are located within the cloud echoes observed by cloud radar. The temporal and spatial correspondences of cloud base observed by cloud radar and ceilometer are poor. (3) Increasing the transmitter power of ceilometer and improving the operational mode of cloud radar can improve the cloud observation abilities. The cloud radar and ceilometer data merging would enhance cloud observation abilities. The research results provide references for usages of cloud radar and ceilometer data and evaluation of observation abilities of cloud radar and ceilometer. They are also helpful in designing operational modes for cloud radar, and promoting the development and application of advanced cloud observation technology.