With increasing flexibility in manufacturing systems, the integrated approach to the planning problems on the shop-floor level has become much more important in recent years. Although adopted process plans may considerably affect the performance of lot streaming, previous research has assumed that parts are to be produced with a fixed process plan throughout the system. This paper addresses the potential improvement in the schedules that can be achieved through integration of process planning and lot streaming decisions. The integrated problem is referred to as flexible lot streaming (FLS). Two integer programming models are developed for FLS problem with and without set-up times. The objective is to minimise the makespan of the schedule generated with consistent sublots. To analyse the effect of integrated approach on the performance of lot streaming, optimal schedules are obtained on a large number of randomly generated test problems. The computational results show that FLS may yield substantial improvement in the makespan for both cases of with and without set-ups. Furthermore, the experiment highlights the impact of speed-up constraints on the computational performance of proposed model.