Over the years, interval-based fuzzy time series forecasting models have garnered significant interest from researchers across various fields. Unlike traditional time series models, the fuzzy time series forecasting models provide flexibility by not imposing stringent assumptions on data characteristics and can be applied to incomplete time series. However, the forecasting accuracy of these models heavily depends on effectively determining the lengths of intervals and output defuzzification rules. Motivated by this challenge, we propose a fuzzy time series forecasting model (FTSFM) based on the index of fuzzy sets, utilizing particle swarm Co-optimization to address these two critical factors. To address the first factor, particle swarm optimization (PSO) is employed to optimize the fuzziness parameters of hedge algebras (HA) and the lengths of intervals in the universe of discourse (UD) of time series data. Additionally, to enhance forecasting accuracy, a novel and more efficient formula for calculating crisp forecasting values based on the index of fuzzy sets is introduced. To evaluate the accuracy of the proposed model, three distinct time series datasets are considered and compared in performance with several recently developed models. The experimental results show that the proposed model achieves better forecasting performance than the comparative models and is more suitable for handling strongly varying data series. Introduce.