Planned lead times are crucial parameters in management of supply networks that continue to be more and more extended with multiple levels of inventory of components and uncertainties. The object of this study is the problem of determining planned lead times in multi-level assembly systems with stochastic lead times of different partners of supply chains. A general probabilistic model with a recursive procedure to calculate all the necessary distributions of probability is proposed. A Branch and Bound algorithm is developed for this model to determine planned order release dates for components at the last level of a BOM which minimize the sum of inventory holding and backlogging costs. Experimental results show the behaviour of the proposed model and optimisation algorithm for different numbers of components at the last level of the BOM and for different numbers of levels and values of holding and backlogging costs. The model and algorithm can be used for assembly contracting in an assembly to order environment under lead time uncertainty.