Mould-repair is an important task in the mould production and depends especially upon the experience of the engineer. It is much full of subjectivity and uncertainty. How to find the knowledge of mould repair and establish the mould repair solving project is an important problem for all the injection mould manufacturers in the industry. Using the Bayesian networks reasoning to help the technician establish the mould-repair solving project was put forward. The mould-repair Bayesian networks (MRBN) reasoning model was established according to the characteristic of mould-repair solving project. The reason control tactic of the process plan system and the reasoning engine of MRBN were designed. The application example about the shrinkage, ill-full injection, gas-obstructed is presented to prove the correctness of the method. The conclusion shows that it can help the engineers analyze the cause of problem well and recommend some more useful solutions.