Wireless mobile networks tend to drop a large portion of packets due to propagation errors rather than congestion. To Improve reliability over noisy wireless channels, wireless networks can employ forward error correction (FEC) techniques. Static FEC algorithms, however, can degrade the performance by poorly matching their overhead to the degree of the underlying channel error, especially when the channel path loss rate fluctuates widely. This paper investigates the benefits of an adaptable FEC mechanism for wireless networks with severe packet loss by analytical analysis or measurements over a real wireless network called sensor network. We show that our adaptive FEC named FECA (FEC-level Adaptation) technique improves the performance by dynamically tuning FEC strength to the current amount of wireless channel loss. We quantify these benefits through a hybrid simulation integrating packet-level simulation with bit-level details and validate that FECA keeps selecting the appropriate FEC-level for a constantly changing wireless channel.