This study uses fuzzy clustering technique to develop a predictive model for interval time series. The proposed model is divided into three phases, with improvements built for each phase. First, overlap distance is used to evaluate the similarity of elements in a universal set of normalized variable data series. The overlap distance is then used to divide the universal set into clusters with an appropriate number. Second, the fuzzy relationship between each element and the clusters is determined based on the fuzzy cluster analysis technique. Third, a new rule is created to interpolate the historical data and forecast the future. The proposed model is detailed for each step and demonstrated using a numerical example. Further- more, this study theoretically demonstrates convergence. Finally, a MATLAB procedure is created for the proposed model which can be easily implemented for practical applications. Several benchmark data series are used to demonstrate practical applications and show the advantages of the proposed model compared to other models.