Monitoring worshiper movements in Islamic prayer and detecting errors is an active field of research due to the importance of the Islamic prayer performed five times a day. The worshiper may make a mistake when performing conditions in prayer by forgetting, or doubt, omissions, disease, etc. This study aimed to detect errors that may occur in the conditions of Islamic prayer at the school of the greatest Imam Abu Hanifah al-Nu'man; by producing a core of errors detection process, an intelligence model able to monitor all prayer actions from starting to end; also, this model able to distinguish the gender of the person performing prayer. An Islamic Hanafi prayer (IHF) dataset was built; To achieve that model, it included both the depth and the skeletal databases using the second version of the Microsoft Kinect sensor, which contains more than 8000 samples. Then, an appropriate spatiotemporal GCN skeletonbased algorithm was used to train and test 4,012 data samples of the IHF skeletal database. Finally, two cross-view and cross-subject evaluations were applied. This study achieved 96.55% and 84.34%, respectively, for each of them.