Abstract: In order to solve the difficult problem of difficult quantification of fracturing effects, taking the hydraulic fracturing of the 14201 working face hard roof in Shenghai Coal Mine as the background, adopting pressure and flow rate fracturing data as clustering indicators, a type of intelligent identification method for hard roof hydraulic fracturing effects based on particle swarm optimization K-means algorithm (PSO-KM) is proposed. This method conducts comparative verification through drilling peeping, the results show that when the fracturing effect is divided into four levels: poor, average, good, and high-quality, the corresponding pressures and flow rates of the cluster center are 12.70 MPa, 0.43 m³/min, 11.56 MPa, 0.54 m³/min, 9.64 MPa, 0.68 m³/min, and 8.09 MPa, 0.87 m³/min, respectively; The peeping results confirm that the identification effect of fracturing effect levels based on clustering algorithm is significant, verifying the effectiveness and reliability of this identification method; In the 14201 working face hard roof, the proportion of high-quality and good grade fracturing sections is as high as 77.48%, and the periodic weighting step distance of the working face is shortened from an average of 50m before fracturing to an average of 28.57 m after fracturing, indicating that the fracturing technical parameters of this working face are reasonable and feasible, effectively reducing the manifestation of roof weighting and achieving safe mining. The research results provide a new method for the intelligent prediction of roof fracturing effects.