Abstract: It is of great importance to sense the roof weighting and roof caving in advance and self-evaluate initial re- sistance and working resistance for promoting the safety,efficiency and intelligent mining level in longwall penal. Based on the scientific principles of intelligent sensing of supports-roof status,this paper used the massive data collected by the electro-hydraulic controlled two-leg shield supports,data mining and working cycle analysis technique to extract the characteristic parameters. Then the supporting characteristics of a single support and the load transformation and distri- bution on the longwall panel considering setting pressure, yielding, and shearer’ s cutting, as well as neighboring shields’ advance were investigated. On the basis of in-depth explanation on the temporal-sequence curve of supporting resistance and the indicating relationship between support and surrounding rocks,the intelligent sensing model was constructed. The study concludes that ① the rated working pressure should not be employed as the main indicator for evaluating the loading capacity of hydraulic support. Parameters,such as initial support pressure,rated working resist- ance and safety valve opening characteristics,would co-determine the bearing characteristics and capacity of hydraulic support. ② By analyzing the massive data of support pressure,the goals of intelligent roof disaster warning,support quality evaluation and fault diagnosis can be achieved. ③ The in-depth analysis and utilization of massive monitored pressure data is the foundation to realize intelligent mining goals such as the self-control of hydraulic support pressure and adaptive adjustment of support parameters.