Abstract: With the advancement of artificial intelligence and cloud computing technologies, drilling analysis and optimization techniques have gained new development opportunities. To fully leverage massive drilling time-series data for real-time downhole condition monitoring, rate-of-penetration optimization recommendations, and drilling risk prevention, this paper develops an intelligent drilling analysis and decision-making system based on cloud-edge collaboration. The system establishes dual closed-loop decision flows through cloud-edge model orchestrated computation and task scheduling: a risk prevention loop ("Edge alert - Cloud analysis - Instruction feedback - Model update") and an optimization loop ("Edge recommendation - Cloud optimization - Solution refinement - Rig execution"). Core innovations include real-time friction torque prediction (accuracy≥91%) via 3D soft-string models integrated with LSTM networks, lost circulation early warning (accuracy>80%) through CNN-LSTM spatiotemporal feature mining, and precise identification of 27 operational phases using a three-tier "Threshold-Intelligent-Dynamic Optimization" model. Validated across 2,000+ wells in Sinopec's northwestern and Zhongyuan fields, delivering comprehensive drilling acceleration and safety management through rig-cloud collaborative decision-making.