Just-In-Time (JIT) is a popular philosophy in many industrial practices. The concept of JIT in early studies concerned with improving operational efficiency and waste minimization. In recent decades, however, JIT principles have also connected to logistics efficiency particularly for distribution of raw materials and finished goods. In the literature, several attempts have been made to optimize JIT logistics networks. On the one hand, most studies have typically focused on deterministic and small-scale problems which have been solved by exact algorithms. On the other hand, when large-scale problems were considered and usually were solved by metaheuristics algorithms, uncertainty sources and fine-tuning of the metaheuristics parameters were generally ignored. In this paper, we develop a mixed-integer linear optimization model to investigate a large-scale JIT logistics problem with 15 different sizes. To deal with different uncertainty sources, the customers demand and suppliers’ capacity as the two main sources of uncertainty in practice are considered as triangular fuzzy parameters. The proposed model aims to minimize total logistics cost including costs of transportation, inventory holding and backorders. A particle swarm optimization algorithm is applied to solve the problem, and its results are then validated by a harmony search algorithm. Both algorithms parameters are tuned using response surface methodology and Taguchi method. Finally, the conclusion and some directions for future research are proposed.