Abstract: Azimuthal gamma measurement has the advantages of flexible implementation and low interpretation complexity, and is widely used in geo-steering decision-making. However, automatic interpretation of azimuthal gamma while drilling has not yet been achieved. A dynamic time warping algorithm is used to identify the formation-penetration which occur when there is a relative angle between the trajectory direction and the target layer while geo-steering, and methods such as piecewise-linear-approximation and Pearson coefficient are introduced to optimize the high computational complexity and singular point problems in the algorithm. The algorithm can use upper and lower gamma log, as well as left and right gamma log, to guide the geo-steering decision adjustment of well inclination and azimuth, and conduct examples respectively. The results show that the improved algorithm significantly reduces computational complexity and meets the real-time geo-steering requirements. Compared to manual recognition, it has higher accuracy. It can not only recognize the formation-penetration features of the trajectory when approaching the boundary, but also accurately identify some interlayer features that are difficult for the naked eye to distinguish, which helps to improve the passive geo-steering caused by shallow gamma detection depth. The application of this method can increase the geo-steering KPI of complex oil and gas reservoirs such as ultra-thin layers and narrow fluvial channels to over 90%. The formation-penetration identification method using dynamic time warping while azimuthal gamma geo-steering has changed the conventional analysis mode that relies on manual experience, providing reference for intelligent application while drilling.