Wireless Sensor Networks (WSNs) consist of low-power sensor nodes with strict constraints on energy, memory, and computational capacity. Yet they must transmit large volumes of data over limited bandwidth and energy budgets. Efficient data compression is essential to prolong node lifetime while minimizing memory and energy overhead. This paper presents MOR-ALDC—a Memory-Optimised Residual Adaptive Lossless Data Compression algorithm tailored for WSN environments. MOR-ALDC enhances the classical ALDC approach by combining residual transformation with entropy-aware canonical Huffman encoding, achieving high compression efficiency with minimal memory use. Extensive evaluations on real-world Intel Lab WSN traces and diverse synthetic signals demonstrate MOR-ALDC’s effectiveness. It achieves compression ratios of up to 8.53× and energy savings of 88.27% on synthetic data, and 5.79× compression with 82.73% energy reduction on real traces. Compared to general-purpose codecs such as zlib, LZ4, and Zstandard, MOR-ALDC delivers superior performance with significantly lower RAM usage, as low as 3–16 KB. Its lightweight, lossless design makes it ideal for next-generation WSNs requiring long-term, high-fidelity data acquisition on resource-constrained hardware.