Process discovery is a process to observe behaviour in the event log and to build a model for the next process. In addition, it is an important process because its strength to predict the time, and cost. After constructing the model, the model has to be set into some parts by using decomposed process algorithm. So that, the result will be easier to be analyzed. A decomposed process can implementation inductive miner algorithm. However, decomposed using inductive miner have limited relation in process. To overcome this problem, this paper proposed decomposed model by using Linear Temporal Logic (LTL) to find the rule and build the process model automatically without constructing from the first step and also have many notations to formalize relation of activity. In addition, LTL is a method to build the rule and check the workflow of the process logs whether the logs have the parallel process. So that, by using the proposed method LTL will be used for getting a process model in less time with average time 1 second and more accurate result.