To study the effect of micro (mi)RNA on cellular proliferation induced by hepatitis B x protein, HBx, in human liver cells and to investigate the underlying molecular mechanism of this cancer-related effect. The human L02 hepatocyte cell line was stably transfected with HBx (L02/HBx) or an HBx mutant (L02/HBx-d382) that induces higher levels of cellular proliferation. The differential miRNA expression profiles were determined by microarray analysis and confirmed by real-time PCR. Two miRNAs, miR-338-3p and miR-551b, that were found to be significantly down-regulated in the L02/HBx-d382 cells were selected for further study and transfected individually into cells using the lipofectamine procedure. The cell survival rate was analyzed by MTT assay, and cell cycles were assessed by flow cytometry. Expressions of cyclinD1, cyclinG1, and E2F1 were assessed by real-time PCR and Western blotting. Compared with the microarray miRNA profile of L02/pcDNA3.0 cells, six miRNAs were up-regulated and five miRNAs were down-regulated in the L02/HBx-d382 cells, while four miRNAs were up-regulated and 12 were down-regulated in the L02/HBx cells. The microarray results were consistent with real-time PCR results. Transfection of miR-338-3p and miR-551b significantly inhibited the cell survival rates (P less than 0.001) and induced G0/G1 phase cycle arrest. According to MTT results: for L02/HBx-d382 cells, compared with lipofectamine or non-transfected (NC) controls, the t value of miR-338-3p was 10.402, 9.133 and the t value of miR-551b was 8.763, 7.403; for L02/HBx cells, compared with lipofectamine or NC controls, the t value of miR-338-3p was 9.105, 8.074 and the t value of miR-551b was 7.673, 7.52. According to flow cytometry results: for L02/HBx-d382 cells, compared with lipofectamine or NC controls, the t value of miR-338-3p was 12.173, 11.107 and the t value of miR-551b was 15.364, 13.377; for L02/HBx cells, compared with lipofectamine or NC controls, the t value of miR-338-3p was 15.416, 13.378, and the t value of miR-551b was 13.276, 13.109. The protein levels of cyclinD1, cyclinG1, and E2F1 were significantly reduced by both miR-338-3p and miR-551b ( P less than 0.001). For L02/HBx-d382 cells, compared with lipofectamine or NC controls: E2F1 had t = 11.132, 10.031 and 12.017, 10.973, respectively; cyclinD1 had t = 15.654, 15.013 and 15.447, 14.733, respectively; cyclinG1 had t = 8.017, 7.661 and 7.402, 7.417, respectively. For L02/HBx cells, compared with lipofectamine or NC controls: E2F1 had t = 14.244, 13.331 and 15.022, 14.468, respectively; cyclinD1 had t = 8.695, 8.137 and 7.877, 7.503, respectively; cyclinG1 had t = 7.73, 7.471 and 7.596, 7.41, respectively. In contrast, the mRNA levels for E2F1, cyclinD1, and cylcinG1 showed no significant differences between the miRNA transfected cells and controls. Wild-type HBx and the high proliferation-inducing mutant HBx can influence the miRNA expression profile of L02 cells. HBx down-regulates miR-338-3p and miR-551b in L02 cells, and the high proliferation-inducing mutant has a more robust effect. The mechanism of miR-338-3p- or miR-551b-mediated cell growth inhibition appears to be related to the direct modulation of cyclinD1, cyclinG1, and E2F1.