The pika (Ochotona curzoniae) hole is an important landscape feature in the Tibetan Plateau (TP) grasslands, and it indicates grassland degradation levels due to the destruction caused by pika burrowing activities on grasslands. However, no studies have ever explored landscape patterns of pika holes and their effects on adjacent vegetation coverage. Taking meadow grasslands in Northern Tibet as an example, this study gathered unmanned aerial vehicle (UAV) images and explored landscape patterns of pika holes and their effects on grass coverage in the surroundings. The performances of two classification methods, including the decision tree classification based on Fully Constrained Least Squares (FDC) and the object-oriented classification (OBC) were compared in recognizing sizes and shapes of pika holes. The results showed that: (1) The object-oriented classification exhibits higher classification accuracy in identifying pika holes. (2) The average size of pika holes in the study area is 0.01 m2 and they exhibit clustered distribution patterns. The average distance between any two nearest pika hole patches is 0.79 m. (3) It presents a significant quadratic relationship between the number of pika holes and grass coverage. (4) The average effective distance of pika holes on the surrounding grass coverage is 20 cm. The findings of this study can provide guidelines for pika control and improve grassland management on the TP.