It is clear from worldwide research that micro chemical processes (MCPs) offer a unique approach to the spatial and temporal control of chemical reactions. The well-known advantages of MCPs are often counterbalanced by serious faults such as channel blockage and catalyst deterioration. To realize stable long-term operation of MCPs, it is necessary to develop a monitoring system that can detect and diagnose these faults. In this work, a physical model-based process monitoring system for a tubular microreactor is developed. A state space model is derived by using the orthogonal collocation method, and the extended Kalman filter is used as an observer. The optimal sensor locations are determined so that unknown parameters such as catalyst effectiveness can be estimated most accurately. In addition, the validity of the conventional observability measures in solving the sensor location problems of MCPs is assessed. Finally, operation policies and control structures for micro chemical plants with an external numbering-up structure are also investigated. Two types of operation policies, total flow control and pressure drop control, are compared from the viewpoint of flow uniformity when blockage occurs.