Sri Lanka's rice systems are subject to low yield events that threaten national food security. Extreme climate events during the cropping season are the main cause, but whether human-induced climate change has contributed to low yield events is an open question. Here, we present an impact attribution analysis that quantifies the effect of climate change to the average yield in 1981-2019 and the low yield event that occurred in 2017 using factual and counterfactual climate model simulations as inputs to three process-based crop models, DSSAT, APSIM and CYGMA. All of the crop models consistently show that climate change has decreased average yield by - 4.99% to - 0.20%, compared to that without climate change. However, the effect of climate change to the 2017 event is mixed in the sign across the crop models. When using a multi-model ensemble average (MME) of the three crop models, a significant negative impact on the Yala season is detected. The large uncertainties associated with the use of different crop models also make it inconclusive whether the 2017-level low yield events would become more frequent and severer by mid-century (2031-2069) under projected climates than under the present-day climate. The same result was derived even when MME is used. These results underscore the need for improved impact attribution to inform climate negotiations on the development of climate-resilient agri-food systems in low-income countries through the Loss and Damage mechanism.