Power management of Wireless network interface cards (WNIC) plays a significant role in the development of mobile and pervasive computing. However, current WNIC power management often substantially degrades performance or even increases overall energy consumption when used with interactive applications. This paper proposes a WNIC Energy management protocol for latency-sensitive interactive applications(ELIA). ELIA exploits features of user interactions for wireless network energy conservation. The effective user interaction analysis techniques recognize complex sequences of mouse clicks or keystrokes, and provide valuable hints to anticipate underlying network access patterns. The prediction algorithm uses dynamic time wrapping techniques to predict future network access based on those interaction hints, and then adapts power mode of the WNIC according to the prediction. Real-world evaluation results show that ELIA achieves a better balance of the competing goals of energy conservation, interactive performance, and application quality compared with state-of-the-art power management methods.