Dry eye disease (DED) currently has no satisfactory treatment partly because of the lack of informative animal models. We evaluated the anti-inflammatory phosphosulindac (PS) for the treatment of DED using a new rabbit model of DED based on the concanavalin A (Con A) acute DED model: we injected all lacrimal glands with Con A weekly under ultrasound guidance, which prolonged DED to >3 weeks, and thoroughly assessed efficacy with tear break-up time (TBUT), tear osmolarity, Schirmer test, and tear lactoferrin levels. Rabbits with DED (n = 8–10 eyes per group) were treated topically with PS or vehicle 3×/day for 21days. PS restored TBUT, tear osmolarity, and lactoferrin levels (P< 0.0001–0.04) to normal but did not significantly improve the results of the Schirmer test. PS showed no side effects and was much more efficacious than cyclosporine or lifitegrast. In the cornea, PS suppressed the activation of nuclear factor kappa-B, the levels of transforming growth factor beta, interleukin-1 beta, interleukin-6, and interleukin-8, and the levels of matrix metalloproteinase (MMP)-1 and MMP-9, and MMP activity. Levels of prostaglandin E2 (PGE2) in tears and cornea were preserved in PS-treated rabbits. Ketorolac and diclofenac, two ophthalmic NSAIDs causing corneal melt, nearly completely suppressed PGE2 levels but had no effect on MMPs. The effects of PS on PGE2 and MMPs likely account for its apparent ocular safety. Our results establish an animal model for acute and chronic DED suitable for drug efficacy studies and indicate that PS merits evaluation for DED.