In the present scenario, the concept of fuzzy set theory is an interesting topic among researchers for forecasting the uncertainty in the time series (TS) data. Subject to the uncertainty issues, plenty of forecasting models have been developed with higher forecasting accuracy. However, there are still some areas where it needs to improve the forecasting model for achieving better accuracy. Concerning this, the current study works on two issues: (1) a length-based discretization approach is applied to find out the universe of discourse as well as the number of intervals for the TS data and (2) the hybrid Jaya-Pi-Sigma neural network (JPSNN) model is considered to model the fuzzy logical relationship, which is further contemplated for forecasting the TS data. The Jaya optimization method is used to optimize the weights of the Pi-Sigma model. To assess the forecasting credibility, the executed result was studied using ten TS data with three profound forecasting models. The outcome of various models using different parameters shows the outperformance of the JPSNN model compared to the comparative models.