Recently, more and more metal-organic frameworks (MOFs) have been directly used as anodic materials in lithium-ion batteries, but judicious design or choice of MOFs is still challenging for lack of structural-property knowledge. In this article we propose a pillared-layer strategy to achieve improved Li-storage performance. Four Mn(II) and Co(II) MOFs with mixed azide and carboxylate ligands were studied to illustrate the strategy. In these 3D MOFs, layers (1, 3, and 4) or chains (2) with short bridges are linked by long organic spacers. All the MOFs show very high lithiation capacity (1170-1400 mA h g at 100 mA g) in the first cycle owing to the rich insertion sites arising from the azide ion and the aromatic ligands. After the formation cycles, the reversible capacities of the anodes from 1, 3, and 4 are kept at a high level (580-595 mA h g) with good rate and cycling performance, while the anode from 2 undergoes a dramatic drop in capacity. All the MOFs lose the crystallinity after the first cycle. While the amorphization of the chain-based framework of 2 leads to major irreversible deposit of Li ions, the amorphous phases derived from the pillared-layer frameworks of 1, 3, and 4 still retain rich accessible space for reversible insertion and diffusion of active Li ions. Consistent with the analysis, electrochemical impedance spectra revealed that the pillared-layer MOFs led to significantly smaller charge-transfer resistances than 2. Soft X-ray absorption spectroscopy suggested that no metal conversion is involved in the lithiation process, consistent with the fact that the isomorphous Co(II) (3) and Mn(II) (4) MOFs are quite similar in anodic performance.