A rule-based object/task modelling approach is proposed which is characterizes by specifying object behaviors and domain rules in terms of object-oriented logic programming and specifying tasks and meta-rules in terms of network-oriented formalism. in addition the concepts of associations virtual objects, multiple level integrity control and net expressions are introduced the object-oriented logic programming system is extended for supporting the sematic modelling and an explicit control knowledge representation mechanism is developed this approach may be viewed as a step to the integration of object-oriented programing logic programming semantic modelling and event modelling and to the combination of forward chaining and backward chaining techniques therefore it can provide complementary benefits in deductive query support integrity control explicit control knowledge representation and intelligent user interface and enhance the flexibility and extendability of knowledge based systems to accommodate applications in multiple domains towards a generalized rule-based management of data actions an operational schemes this approach is being designed and partially implemented on top system g [Chen 85b] on a vax computer.