Moxibustion is effective for low back pain (LBP), and inflammatory cytokines may play an important role in the mechanism of moxibustion treatment. The purpose of this meta-analysis was to explore the mechanism of moxibustion in LBP in terms of inflammatory cytokines. We searched China National Knowledge Infrastructure, Wanfang database, Cochrane Central Register of Controlled Trials, Ovid MEDLINE, Embase, PubMed, and Web of Science to identify eligible randomized controlled trials (RCTs). There was no restriction on the publication date. Thirty RCTs measuring interleukin (IL-) 1, IL-1β, IL-6, IL-12, IL-17, IL-23, and tumor necrosis factor (TNF-) α were included in this meta-analysis. Compared to controls: single moxibustion could effectively decrease levels IL-6 and IL-23 (SMD, -0.71, 95% CI: -1.25 to -0.17, p = 0.01; SMD, -1.61, 95% CI: -2.20 to -1.03, p < 0.01, respectively); combined moxibustion had significant effects on IL-1, IL-1β, IL-6, IL-12, IL-17, and TNF-α (p < 0.05). Overall, for LBP, single or combined moxibustion could effectively down-regulate levels of pro-inflammatory cytokines (p = 0.007 and p < 0.00001, respectively). For safety of moxibustion, the incidence rate of side effects was similar to that of controls (RD, -0.01, 95% CI: -0.02 to 0.01, p = 0.59). Sensitivity analysis showed that the pooled estimates were robust, and publication bias analysis showed there was a significant small study effect (Egger's test p = 0.0000). High statistical heterogeneity existed between included RCTs, meta-regression showed there was no potential factor explaining the source of heterogeneity. For LBP, moxibustion can effectively decrease levels of IL-1, IL-1β, IL-6, IL-12, IL-17, IL-23, and TNF-α to achieve analgesia. Because the side effects of moxibustion are transient, it is relatively safe for clinical use. However, based on high heterogeneity in this meta-analysis, rigorously designed RCTs are required to further confirm the results in this review.