The job shop scheduling literature deals with problems characterized by a fixed linear process plan for each job: it is assumed that the process planning problem has been solved before scheduling, and no flexibility in the process plan is considered. Our aim here is to propose a solution approach for a joint process plan selection and job shop scheduling problem, taking both operations cost and makespan into account within a multi-objective framework. Due to the complexity of the problem, a two-phase hierarchical method is proposed. In the first phase, a relaxed version of the problem is solved, yielding an approximation of the set of efficient process plans with respect to cost and load balancing objectives. Each process plan is then considered and the corresponding scheduling problem is solved by tabu search; the process plan selection is improved by a two-level hierarchical tabu search algorithm.