Reliable and effective forward error correction is challenging, especially in a heterogeneous transmission environment, due to the differences in computation capabilities of end devices. Using a complex code to achieve reliable communication leads to high computation costs and long decoding delay at devices with low computing power. Fulcrum codes, a variation of random linear network coding (RLNC), addresses that problem by combining two codes of large and small Galois field sizes. Thus, a Fulcrum decoder can decode using either field size. However, state-of-the-art Fulcrum decoders select and operate on a predetermined Galois field throughout the transmission session regardless of current states of received packets. We propose an inclusive and adaptive decoding process that decide when to operate on which Galois field in accordance with the computational capabilities and varying channel environments. Our comprehensive evaluation shows that the proposed adaptive decoding significantly reduces the computation complexity at end devices while simultaneously maintaining a high decoding probability with substantially low overhead and decoding delay.