Objective: To examine the cost-effectiveness of self-managed computerised word finding therapy as an add-on to usual care for people with aphasia post-stroke. Design: Cost-effectiveness modelling over a life-time period, taking a UK National Health Service (NHS) and personal social service perspective. Setting: Based on the Big CACTUS randomised controlled trial, conducted in 21 UK NHS speech and language therapy departments. Participants: Big CACTUS included 278 people with long-standing aphasia post-stroke. Interventions: Computerised word finding therapy plus usual care; usual care alone; usual care plus attention control. Main measures: Incremental cost-effectiveness ratios (ICER) were calculated, comparing the cost per quality adjusted life year (QALY) gained for each intervention. Credible intervals (CrI) for costs and QALYs, and probabilities of cost-effectiveness, were obtained using probabilistic sensitivity analysis. Subgroup and scenario analyses investigated cost-effectiveness in different subsets of the population, and the sensitivity of results to key model inputs. Results: Adding computerised word finding therapy to usual care had an ICER of £42,686 per QALY gained compared with usual care alone (incremental QALY gain: 0.02 per patient (95% CrI: −0.05 to 0.10); incremental costs: £732.73 per patient (95% CrI: £674.23 to £798.05)). ICERs for subgroups with mild or moderate word finding difficulties were £22,371 and £21,262 per QALY gained respectively. Conclusion: Computerised word finding therapy represents a low cost add-on to usual care, but QALY gains and estimates of cost-effectiveness are uncertain. Computerised therapy is more likely to be cost-effective for people with mild or moderate, as opposed to severe, word finding difficulties.