Network coding (NC) can reduce the number of packet retransmissions and improve network reliability in lossy networks by combining packets in generations and transmitting these coded packets. Although corrupted coded packets are discarded at the receiver, they remain valuable as the majority of their data remains error-free, even in the presence of a noisy channel with a bit error rate range of 10-5. Numerous packet recovery strategies have been proposed to tackle this issue; however, when confronted with excessively noisy channel conditions, their recovery speeds experience a substantial decline. This letter presents a new method called Quick partial packet recovery (QPPR). QPPR exploits the algebraic consistency among coded packets to estimate the erroneous parts of packets and utilizes this estimation with channel condition information to recover packets. Through simulations, we demonstrate that QPPR enhances communication throughput in diverse scenarios, achieving up to a 26.64 percent improvement. Additionally, QPPR effectively reduces communication overhead and delay in very noisy channel conditions, showcasing its potential to elevate overall communication performance.