This paper proposes a novel fuzzy forecasting method for forecasting the TAIEX based on optimal partitions of intervals, optimal weights, and particle swarm optimization (PSO) techniques. First, it applies PSO techniques to find optimal intervals and optimal weighting vectors of two-factors second-order fuzzy-trend logical relationship groups (TFSTLRGs) simultaneously using the historical training data (HTD). Then, based on the obtained optimal partitions of intervals in the universe of discourse, it fuzzifies the historical testing data of the main factor (MF) and the secondary factor (SF) into fuzzy sets, respectively. Finally, it chooses a TFSTLRG to perform the forecasting based on the obtained optimal weighting vector of the chosen TFSTLRG. The advantage of the proposed fuzzy forecasting method is that better forecasting accuracy rates are obtained for the forecasting of the TAIEX.