Background: Clinical trials of Attention Bias Modification for depressive symptoms have consistently produced small effect sizes and mixed results. Therefore, identifying patient characteristics that can improve the efficacy has been called for. Residual depressive symptoms have been shown to increase the chance of relapse and treating this group of patients may serve to reduce the risk. In this analysis we examined whether baseline attention bias moderates the effect of Attention Bias Modification, the intended mechanism of change of the intervention. Methods: The analysis was based on data from a randomized controlled trial with participants with a history of depression (N = 301) who were randomized to receive two daily sessions of either Attention Bias Modification or a sham treatment for 14 days. A response-based attention bias score at baseline was calculated, and a moderator analysis was run at post-intervention and 1-month follow-up measured by change in Hamilton Depression Rating Scale (HDRS) and Beck Depression Inventory II (BDI II), respectively. Results: Baseline attention bias did not moderate the effect of Attention Bias Modification at post-intervention (HDRS F(1,295) = .303, p = .582, BDI II F(1,295) = 1.1567, p = .283: or at 1-month follow-up (HDRS F(1,265) = .000, p = .247, BDI II F(1,267) = .021, p = .885). Conclusion: Even though Attention Bias Modification targets attention bias, the baseline score did not significantly moderate the effect of Attention Bias Modification on any of the time points or depression measures. Consequently, baseline attentional bias does not seem to be a useful variable for identifying patients with subclinical depressive symptoms more likely to benefit from ABM.