In order to reduce mechanism end errors and enhance robot motion accuracy，the main error sources of the parallel robot and the limitations of conventional error compensation are discussed for the 6-DOF precision parallel robot designed. An error compensation method based on back propagation (BP) neural networks of articulatory space

by measuring the end pose

is presented in local workspace of precision positioning. Model of BP Network and data sample of error compensation are established

and the data sample is standardized. By the experiment

the node number of hidden layer is achieved. In order to improvement the generalization performance

the overfitting is prevented in the network training. After error compensation

the positioning error and the orientation error reduce by 80% and 60% respectively. The experimental results show that the error compensation

based on BP neural networks of articulatory space

has an obvious effect

which satisfies the accuracy requirements of the precision parallel robot.