Process planning is one of the key activities for product design and manufacturing. Process plans developed in this activity fulfil several purposes: manufacturability evaluation of product design, cost justification of required processes, process and geometry data for NC programming, and process specification for resource planning, scheduling and actual manufacturing. Ability to consider and evaluate alternative plans during design and/or scheduling allows for better product design, lower cost and more efficient manufacturing. In this paper we describe the process plan network, a hierarchical model which allows other functions to select the most appropriate alternative representation for given conditions. The procedure for generation of the process plan network is explained and two algorithms for the selection of process plans in various stages of product development are presented. Several experimental results are shown. The process plan network consists of several levels: feature level, process level, tool direction level, machine level and plan level. Generation of network starts at the bottom level, from features and by considering alternative processes, tools and machines and clustering of processes into manufacturing activities the network is build toward its top. During the network generation a large set of ordering constraints between features, tools and machines is considered in order to produce set of alternative but feasible process plans. The selection of process plan from the network is performed according to some efficiency criteria. Two optimization algorithms are applied: space search and network optimization. The method for simultaneous network generation and plan optimization is also implemented. The efficiency of algorithms is described on several examples.