Combinational chemistry is an efficient technique to find materials with novel properties by synthesizing and screening a large number of compounds in a short time. Recently, we introduced the concept of combinational approach into computational chemistry and proposed a novel approach, "combinatorial computational chemistry". In the present study, we applied combinatorial cornputational chemistry to investigate the structural properties of lithium transition metal oxides, LiMO2 (M = 3d transitional metal), with a layered rocksalt structure. LiMO2 is a promising material as positive electrodes in rechargeable lithium batteries. Density functional calculations on periodic models were performed to investigate the structural properties of LiCoO2, LiNiO2, and doped LiNiO2, revealing that the poor charge-discharge cyclic reversibility of LiNiO2 resulted from the large change in the structure due to the difference in the bond length between Ni3+-O and Ni4+-O. The analysis of the structural properties of Li0.66Ni0.5Me0.5O2 (Me = dopant) revealed that doping with Co decreased the change in the structure of LiNiO2 during cycling. Doping of Ni with Al was also found to stabilize LiNiO2. (C) 2002 Elsevier Science B.V. All rights reserved.