The severe nature of underwater channel poses a great challenge for prolonging underwater acoustic sensor networks (UASNs) lifetime and achieving a reliable communication performance. Traditional approaches to improve the reliability such as automatic repeat request (ARQ) negatively affect the network lifetime (NL) due to energy dissipation caused by ARQ retransmission. A forward error correction (FEC) technique called fountain codes (FCs) can solve the energy efficiency problem of ARQ by transmitting both the original packet and some redundant packets to ensure a targeted reliability with few or no retransmissions. In this paper, we investigate performances of both traditional ARQ-and FC-based FECmethods in terms ofNL, end-to-end delay, energy consumption, and frame error rate (FER) for UASNs. In this context, we abstract energy dissipation characteristics of conventional ARQ-and FC-based FECmethod at the link-layer. We propose an integer linear programming (ILP) framework that maximizes the NL, which is operated on top of developed link-layer energy consumption models. Our results reveal that FC-based FEC methods can prolong the NL at a minimum of 16% while end-to-end delay, energy consumption, and FER can be reduced at least by 11%, 14%, and 9% as compared to classical ARQ, respectively.