This paper investigates the synchronization problem of delayed neural networks (DNNs). First, we introduce an intelligent hybrid eventtriggered scheme (IHETS) that allows sensors to wait within a specified time range after triggering measurements, thereby reducing network load and avoiding Zeno behavior. Then, a semi-looped-functional approach with relaxed constraints is adopted, which eliminates the need for strict requirements on negative definite derivatives. Moreover, we introduce a novel series of less conservative linear matrix inequalities (LMIs) to guarantee sufficient conditions for the existence of asymptotic synchronization controllers, considering practical engineering factors such as actuator failures. Finally, the effectiveness of this approach is validated through a numerical example.