Data compression can efficiently utilize the storage capacity. Especially for mobile devices, they have limited resources such as a CPU (Central Processing Unit) operation clock and a power source, so hardware designed compression method is required. Most adaptive dictionary-based compression methods have their root in Lempel-Ziv algorithm. LZ4 is one of the fastest compression algorithms. In this paper, we propose an advanced algorithm and hardware architecture with improved compression ratio and speed. In order to achieve a higher compression ratio, the proposed algorithm has a variable length format while LZ4 has a fixed length format. Experimental results show that the proposed architecture can achieve compression throughput up to 3.84Gbps and compression ratio up to 4. In this way, our hardware-based architecture improves memory performance and battery life of mobile devices with low power consumption.