The work presented in this paper investigates the effects of lot size on the performances of job shop production systems. This subject remains a very real problem in discrete production shops, especially when a just-in-time production strategy is adopted and small lot sizes and high throughput are needed. To respond to customers' requirements at the right time and in appropriate quantities, this strategy requires, greater flexibility, faster set-up or changeover time, different product batches, limited lot sizes and mainly short manufacturing lead times from production systems. The originality of this investigation resides in the method used to analyse the impact of the lot sizes and set-up times on the design and optimisation of a job shop or intermittent production system. Knowing the product types, product routings, process times, machine types and the demand forecasting, we seek an optimal lot size that enables an optimisation of the production design system. Thus, we study in this article the influences of the lot sizes: first on single-product manufacturing without set-up time and second on multi-product manufacturing with crossing routings and changeover times. For the two cases, this study is investigated at variable and constant flow or demand orders. To determine optimal lot size, a performance criterion was defined, based on: total production, average production lead time and average activity rate of the system. The model and tests are being carried out through simulation with the general language tool “Siman”. The graphical representations of results concerning performance measures such that manufacturing lead times, total production, system utilisation and the above defined performance criterion show some very interesting points. Whatever the studied case, an optimal lot size could be defined. From this optimal lot size value, when lot size increases, lead times and work-in-process inventories increase, and total production and system activity rates decrease. On the other hand, if lot size decreases from the optimal value, total production and average utilisation quickly diminish, causing longer lead times and increased work-in-process inventories. In the last step, this optimal lot size allowed us to balance the different work centres to have a well-designed physical system.