Perovskite Li
3xLa
2/3–xTiO
3 (LLTO) has been investigated as a Li-ion solid electrolyte material and has attracted significant attention due to its wide operating voltage range. Polycrystalline and grain boundaries (GBs) are a common structural motif found in ceramic oxides. So, GBs can have a significant influence on the material properties. Here, we conduct a molecular dynamics (MD) study that quantifies the effect of LLTO GBs on Li-ion transport. We examine six types of LLTO GBs, including P-Σ5(210), P-Σ5(310), P-Σ13(510) in the Li-poor phase and R-Σ5(210), R-Σ5(310), R-Σ13(510) in the Li-rich phase. We also consider the LLTO bulk for comparison. The results show that the grain boundary formation energies of the six GBs are all below 1.30 J/m
2, indicating the presence of a high concentration of GBs in polycrystalline LLTO. It is likely to find a highest concentration of Σ5(210) GB due to its lowest formation energy (1.00 J/m
2 for P-Σ5(210) and 0.89 J/m
2 for R-Σ5(210)). Compared with Li
+ ions in the bulk LLTO, Li
+ ions in the six GBs exhibit a lower mean squared displacement (MSD), a smaller migration energy barrier and a lower ionic conductivity. These results confirm that the LLTO GBs hinder Li
+ transport. For bulk LLTO, the Li
+ migration barrier is determined to be 0.30 eV (Li-poor phase) and 0.26 eV (Li-rich phase). In comparison, the migration barrier of LLTO GBs exhibits a slight decrease in Li-poor phase (0.32–0.37 eV) and Li-rich phase (0.27–0.31 eV). The computed Li-ion conductivities of the six GBs are 1 to 2 orders of magnitude lower than those of the corresponding bulk counterparts. Of the six GBs, P-Σ13(510) exhibits the highest Li
+ conductivity of 4.76×10
–5 S/cm in the Li-poor phase, whereas R-Σ5(310) shows the maximum Li
+ conductivity of 1.31×10
–3 S/cm in the Li-rich phase. Furthermore, the peak Li
+ conductivity in the Li-rich phase is much higher than that in the Li-poor phase. In addition, Li
+ transport perpendicular to the GB (i.e. from grain to grain) is more hindered than that along the GB. Nevertheless, the Li
+ diffusion can be improved by increasing the Li content in the GB region. The Li
+ diffusion maps can be visualized by analyzing the Li
+ trajectories of the MD simulations. We find that Li
+ transport is first restricted to the GB region, then gradually turns to the bulk region, and finally forms a two-dimensional diffusion path similar to that of the LLTO bulk. Furthermore, the Li
+ diffusion strongly depends on the distribution of O ions in LLTO GBs. For example, in the Li-poor-phase P-Σ5(310) GB, the number of O ions in the GB region is greater than that in the bulk region, which indicates that a Li-O attractive interaction in the GB region is stronger, thereby hindering the formation and transport of Li
+ to the bulk region. Overall, these atomic-scale insights deepen our understanding of LLTO GBs and their influence on Li
+ transport.