This paper evaluates the performance of multiple-load automated guided vehicles (AGVs) under different conditions, and compares those results against the ones obtained with a single-load vehicle system. We design and conduct a full factorial analysis of variance on several response variables, namely flowtime, production completion, work-in-process, AGVs loaded utilization, and AGVs total utilization. We first run an experiment where load pickup and drop-off rules, vehicle fleet size, shop loading, and buffer capacity are the factors to be analyzed. In addition, we chose the best set of rules for the multiple-load vehicle system and compared their performance against the unit-load system. We found that the FSNS-a (First-Serve-Next-Station with priority to parts arriving to a queue) pickup rule performs best for a multiple-load vehicle, in terms of flowtime, WIP (Work In Process,) and AGVs total utilization. Only one performance measure was sensitive to drop-off rules, namely loaded AGVs utilization. The ND (Nearest Destination) drop-off rule minimizes this response variable. When we compared the multiple-load AGV system using the FSNS-a and ND rules against a unit-load system, we found out that the multiple-load system with only three vehicles performed better than the unit-load system with a fleet of four units. Larger buffer capacities were useful on the multiple-load system working at higher shop loading. Besides producing a contraction on flowtime, WIP inventory was also reduced. Results on the unit-load system did not improve when buffers are augmented.