Summary The conclusions of pairwise meta‐analyses of interventions for actinic keratosis (AK) are limited due to the lack of direct comparison between some interventions. Consequently, we performed a network meta‐analysis for eight treatments 5‐aminolaevulinic acid (ALA)‐photodynamic therapy (PDT), cryotherapy, diclofenac 3% in 2·5% hyaluronic acid (DCF/HA), 5‐fluorouracil (5‐FU) 0·5% or 5·0%, imiquimod (IMI) 5%, ingenol mebutate (IMB) 0·015–0·05%, methyl aminolaevulinate (MAL)‐PDT and placebo/vehicle (including placebo‐PDT) to determine their relative efficacies. As part of a prior Cochrane systematic review, different databases and grey literature were searched for randomized controlled trials up to April 2012. The inclusion criteria were parallel‐group studies with nonimmunosuppressed participants: (i) reporting ‘participant complete clearance’ and (ii) comparing at least two of the interventions. Thirty‐two publications met the criteria and they included the following number of individual or pooled studies (n) and total number of participants (N) for the different interventions: 5‐FU 0·5% (n = 4, N = 169), 5‐FU 5·0% (n = 2, N = 44), ALA‐PDT (n = 6, N = 739), cryotherapy (n = 2, N = 174), DCF/HA (n = 5, N = 299), IMI (n = 14, N = 1411), IMB (n = 3, N = 560), MAL‐PDT (n = 7, N = 557) and placebo (n = 32, N = 2520). Network analyses using a random‐effects Bayesian model were carried out with the software ADDIS v1.16.1. The interventions were ranked as follows based on calculated probabilities and odd ratios: 5‐FU > ALA‐PDT ≈ IMI ≈ IMB ≈ MAL‐PDT > cryotherapy > DCF/HA > placebo. This efficacy ranking was obtained based on the current available data on ‘participant complete clearance’ from randomized controlled trials and the analysis model used. However, several other factors should also be considered when prescribing a treatment for AK. What's already known about this topic? Several interventions are effective for the treatment of actinic keratosis (AK). Due to the lack of direct comparisons between these interventions, their relative efficacy is unknown. What does this study add? Using network meta‐analysis, the relative efficacy of eight main interventions for AK was determined based on the outcome ‘participant complete clearance’. With the exception of ingenol mebutate, the relative efficacy of the interventions was independent of the anatomical location of the lesions.