The recently increased seismicity in Oklahoma provides a rich catalog that delineates unmapped fault system. In this project, I identify 88 clusters with 20 or more earthquakes using automatic nearest-neighbor approach. It is assumed that each cluster nucleates on a preexisting fault segment; I manually pick the fault geometry and use a spectral decomposition method to get fault strike, dip and planarity. These seismogenic faults are steeply dipping (>80°) and the dominant strikes are 50°~60° and 100°~110°. To study the fault orientation to regional stress field, I invert the stress state from focal mechanism solutions using a linear stress inversion method. Most areas in Oklahoma show strike-slip faulting regime, while the north and northwest areas show a transtensional regime. By comparing the seismogenic fault orientation to regional stress orientation, I find that most newly reactivated seismogenic faults are optimally or moderately optimally oriented, thus posing a potential earthquake hazard. I show a case study of seismogenic fault in Woodward, Oklahoma. Three fault segments are picked from seismicity and both the main strike-slip fault and shallower normal fault are optimally oriented to regional stress field. I analyze the reactivation process of the fault system using seismicity migration and Coulomb Stress interaction. Finally, I search for dynamic triggering evidence in Oklahoma by both catalog search and waveform search. I find triggered events in Woodward cluster, Oklahoma by May 12th, 2015 M7.3 Nepal earthquake. The triggering stress threshold is as low as 1kPa, suggesting the faults were critically loaded at the mainshock time.