Elongation of underwater acoustic sensor networks (UASNs) lifetime is possible by optimizing the size of the data packet as well as transmission power. While large data packets are sensitive to packet errors frequently occurring in underwater environment, data fragmentation reduces packet errors. However, with this method more data packets are injected to the network. On the other hand, increasing the transmission power is another way to reduce packet errors. However, since UASN nodes consume excessive amount of energy for transmission (in order of Watts), this solution also reduces energy efficiency. In this work, an integer linear programming (ILP) method is proposed that maximizes the network lifetime by jointly optimizing the data packet size and transmission power. The effects of data fragmentation on network lifetime are quantitatively observed in this study.