Abstract: As the high-humidity， dusty and complex environment leads to the distortion of 3D laser detection results for goafs in the underground metal mines， a device is developed to simulate the complex environment of goafs and to carry out 64 groups of point cloud error analysis and modification experiments with different dust concentrations and relative humidity. It could be observed that with the increase of dust concentration and relative humidity，the average error ratio of point cloud data showed an "S" trend. In particular，under the condition of 30.0mg/m³≤ρ≤85.0mg/m³ and 76.0%≤w≤85.0%，the error growth rate of point cloud data increases exponentially. It can be concluded that the ρ ranging from 0 to 5.6mg/m³ and the w ranging from 0 to 49.0% are favorable for the laser detection of goaf. In addition，the error correction formulas of point cloud are put forward for high humidity， dusty and complex environment. After the application of test results to a gold mine in Fujian Province，the revised parameters such as the elevation， the volume and the exposed area of the goaf are more practical，which proves the accuracy of detection data.

Key words: complex goaf, three-dimensional laser detection, dust concentration, relative humidity, error correction