The article describes a structured approach to the collection and verification of data on the process of blast borehole drilling in the formation of geostructural block models. When using methods and means of machine learning to create block models based on interpreted drilling data, an important aspect is the accuracy of the input data for the learning process compared with the geological and structural information about the rock mass. To identify the real response of drilling to the variation of rocks and the structure of the rock, the article provides comparison of data on the drilling process, data on specific energy intensity and data on the analysis of the changeability of the drilling in the rock when entering various geologic zones and fracture zones. To identify the real state of the rock, its variability and response of drilling data, a comparison of the slope of the ledge after the contour explosion is made in conjunction with the information on the specific energy consumption of the discontinent rows passed in blocks. The results of the endoscopic of the main explosive boreholes for the delay of telemetry data with comparison with the specific energy of drilling are presented. The analysis of the research obtained during the studies showed that the measurement system during drilling gives a real response to changes in array, such as breed transitions, the presence of fracture and disorders, etc. For a more accurate assessment of the nature of the appearance of jumps of values, it is possible to compare the specific consumption energy intensity of drilling and data behavior during MWD drilling.