OBJECTIVE: To compare the incremental cost-utility ratio (ICUR) of latanoprost and travoprost versus timolol in France. METHOD: The probability to develop a visual field defect (VFD) was issued from a double-masked double-dummy, phase III multi-centre clinical trial comparing travoprost 0.004% od, latanoprost 0.005% od and timolol 0.5% bid, using the 2 discriminant functions (stable and progressive patients) published by Stewart, 1993. A Markov model was constructed to reproduce the cost and utility of a patient treated over five years, including the following states: VFD and no VFD. Utility was derived from Brown (2000), based on visual acuity. Direct medical and indirect costs were estimated from a five-year retrospective patient chart analysis. Both costs and outcomes were discounted at a 5% rate. The economic perspective was the one of Society. Sensitivity analysis was performed on the 25th to 75th percentile range of the discriminant empirical distribution function. RESULTS: The ICUR of latanoprost over timolol was €20,327/QALY and €15,374 for travoprost. Results without discounting were similar, €20,150 and €15,196, respectively. At the 25th percentile of the discriminant function, the ICUR were €58,250 and €38,428 respectively and €11,262 and €11,639 at the 75th percentile. At the median of the distribution function, travoprost in comparison to latanoprost brings 0.01 QALY at a €51 additional cost. CONCLUSION: According to our model, travoprost is a cost-effective alternative to timolol. The additional cost to be paid for one QALY in comparison to timolol is lower with travoprost than with latanoprost.