Evaluation of the bearing capacity of bedrock is a long-lasting challenge to geotechnical engineering. However, identifying the rock layer distribution using boreholes is usually time-consuming and costly, particularly in areas with a significant depth of overburden soils. The Weathering Grade (WG) of bedrock is critical for identifying bearing capacity, yet the spatial distribution of WG is often overlooked in current studies. This study employs the K-Nearest Neighbors (KNN) method to predict the spatial distribution of WG using data from 127 boreholes in a Macau construction project. The proposed classification model could generate the distribution of WG in 3D space. Validated by the test dataset, the root mean square errors for predicting surfaces of grade III and grade II are only 1.76m and 2.21m, respectively. Moreover, the model achieved an F1 score exceeding 70% in classifying the WG of bedrock. The classification of various WG and elevation of rock layer provide geotechnical engineers with valuable insights into bedrock quality assessment.