Service-oriented architecture SOA is an architectural style where the services are used instead of being physically integrated. The web services participating in a composition may evolve autonomously, making the service composition's behaviour uncertain. Moreover, the unavailability of source code in most of the situations makes prioritisation of the test cases of the web service composition more challenging. Traditional code-based prioritisation techniques are inappropriate. In this paper, a model-based test sequence prioritisation technique uses the coloured Petri nets as a model for web service composition design verification. The effectiveness of the prioritisation technique is validated using the average percentage fault detection APFD metric. The results show that the new proposed approach has high probability to outperform the random ordering on revealing regression faults in a modified web service composition code.