This peper presents a parallel algorithm for optimal control problems. The parallelism is applied at the algorithmic level, which suggests the use of a real-time parallel programming language of high level as for example Ada. The algorithm is formulated as number of rules which control the execution of its main tasks: the "gradient" tasks and the "object function" tasks, to be run concurrently. Simulation experiments are used to compare the performance of the proposed algorithm with that of the corresponding sequential one.