Abstract: Currently, the working resistance data of hydraulic supports are typically acquired periodically through automated data collection. There are two problems to be solved in the identification of working face pressure based on the resistance data of hydraulic supports: one is how to extract the end resistance of cycle data from the vast amount of working resistance data, and the other is how to effectively use these data to accurately reflect the roof pressure state and provide an accurate basis for determining whether the working face is under pressure. To address these problems, this paper proposes a method for identifying the working face pressure based on intelligent detection of coal cutting cycle. Firstly, the coal cutting cycle detection is converted into a binary classification problem, and the feature vector that effectively represents the end of the coal cutting cycle is designed. Random sampling method is applied to deal with the training set imbalance. Secondly, the support vector machine (SVM) classifier is trained to detect the end time of the coal cutting cycle. Then, based on detecting the end times of all coal cutting cycles, the end resistance of cycle data from each support is extracted. Finally, single sequence data reflecting the overall pressure state of the working face is obtained through data fusion, and the working face pressure is identified using the pressure judgment formula. The experimental results based on the working resistance data of hydraulic supports in a certain working face of Buliangou Coal Mine show that the accuracy and F1-score of the coal cutting cycle detection results are 82.6% and 87.5%, which are improvement of 55.0%, and 44.3% respectively compared to the comparison method. The accuracy and F1-score of the pressure identification results are 82.7% and 84.2%, which are 24.0% and 13.8% respectively compared to the comparison method. These results demonstrate that the proposed method is superior to the comparison method in identifying the end resistance of cycle and the working face pressure identification, with significant improvements in accuracy.