Airframe assembly process needs many drilling and countersinking operations. The two main challenges concerning the drilling process are that the holes must fit in the required tolerances in order to ensure the assembly quality, and that the use of drills must be optimal in order to reduce production costs. These two objectives require the implementation of a reliable online monitoring system. A vast amount of work has been done in the field of drilling monitoring. Unfortunately, many methodologies described in these studies are unlikely to leave the labs as they were often considered difficult to implement, unreliable or not viable economically. The use of multisensor systems integrated with intelligent information processing techniques improved reliability and flexibility of tool condition monitoring systems. However, they have mainly been implemented under steady process conditions and in sensor-friendly lab environments, and neither issues about the variability of the operating conditions, nor quality of input data have been tackled. The work presented here is aimed at demonstrating the potential improvements that could be achieved in robust monitoring of drilling operations by using multisensor fusion and associated recent theoretical developments about uncertainty modeling and handling. The monitoring problem will be formalized, and its associated requirements in terms of accuracy and reliability, as well as issues related to industrial implementation will be detailed in order to position the problem. An approach to implement an industrial monitoring system will be proposed that covers the following steps: sensor integration, feature extraction, estimator choice, and monitoring system evaluation. The proposed methodology could be applied to a broader scope of applications, including most complex manufacturing automated operations.