possess binary coding of "1" or "0" on each bit, and their storage capacity is about to reach its maximum. Thus, it is urgent to develop novel storage media with lowpower consumption, high storage density, and long lifespan. Deoxyribonucleic acid (DNA), known as the genetic information carrier, [4,5] has proved to be a promising data storage medium due to its long lifespan, [6][7][8][9][10] sky-high storage density, low energy consumption, and low maintenance cost. [11][12][13] Natural DNA has four nucleotides (A, T, C, and G) and has a storage density of ≈460 EB g −1 , [14] which is much higher than that of traditional storage media. Therefore, the development of new types of DNA molecules to improve information storage is an attractive goal.DNA data storage is mainly affected by the number of nucleotides, the biochemical properties of DNA, and the technical constraints of DNA synthesis and sequencing. [15] Therefore, encoding systems including different functional modules, such as index design and error correction, are required to accommodate the characteristics of the DNA storage channel: i) DNA sequences with acceptable synthetic errors are generally limited to about 250 nucleotides and large data is thus divided into short DNA sequences. Therefore, redundant nucleotides are required to Information storage based on DNA molecules provides a promising solution with advantages of low-energy consumption, high storage efficiency, and long lifespan. However, there are only four natural nucleotides and DNA storage is thus limited by 2 bits per nucleotide. Here, artificial nucleotides into DNA data storage to achieve higher coding efficiency than 2 bits per nucleotide is introduced. To accommodate the characteristics of DNA synthesis and sequencing, two high-reliability encoding systems suitable for four, six, and eight nucleotides, i.e., the RaptorQ-Arithmetic-LZW-RS (RALR) and RaptorQ-Arithmetic-Base64-RS (RABR) systems, are developed. The two concatenated encoding systems realize the advantages of correcting DNA sequence losses, correcting errors within DNA sequences, reducing homopolymers, and controlling specific nucleotide contents. The average coding efficiencies with error correction and without arithmetic compression by the RALR system using four, six, and eight nucleotides reach 1.27, 1.61, and 1.85 bits per nucleotide, respectively. While the average coding efficiencies by the RABR system are up to 1.50, 2.00, and 2.35 bits per nucleotide, respectively. The coding efficiency, versatility, and tunability of the developed artificial DNA systems might provide significant guidance for high-reliability and high-density data storage.