Aluminum-ion batteries, emerging as a promising post-lithium battery solution, have been a subject of increasing research interest. Yet, most existing aluminum-ion research has focused on electrode materials development and synthesis. There has been a lack of fundamental understanding of the electrode processes and thus theoretical guidelines for electrode materials selection and design. In this study, by using density functional theory, we for the first time report a first-principles investigation on the thermodynamic and kinetic properties of aluminum intercalation into two common TiO2 polymorphs, i.e., anatase and rutile. After examining the aluminum intercalation sites, intercalation voltages, storage capacities and aluminum diffusion paths in both cases, we demonstrate that the stable aluminum intercalation site locates at the center of the O-6 octahedral for TiO2 rutile and off center for TiO2 anatase. The maximum achievable Al/Ti ratios for rutile and anatase are 0.34375 and 0.36111, respectively. Although rutile is found to have an aluminum storage capacity slightly higher than anatase, the theoretical specific energy of rutile can reach 20.90 Wh kg(-1), nearly twice as high as anatase (9.84 Wh kg(-1)). Moreover, the diffusion coefficient of aluminum ions in rutile is 10(-9) cm(2) s(-1), significantly higher than that in anatase (10(-20) cm(2) s(-1)). In this regard, TiO2 rutile appears to be a better candidate than anatase as an electrode material for aluminum-ion batteries.