This paper describes an innovative modeling and training framework and an simulator application for micro tunneling machines under heterogeneous gravel and sand soils. From the selective collection of skilled pilots' know-how of a pipe jacking microtunnelling machine in operation, to generate a rule-based system based on grouped rules and states that replicates machine performance. The adjustment of these states and associated rules allows creation, setup and analysis of a realistic functional model for tunneling machines. The system integrates a friendly human machine interface (HMI) that closely replicates real machine's pilot cabinet and allows natural interaction with the implemented inference engine through the simulated control panels. Additionally, the framework allows the training of tunneling machine's operators by simulation and subse- quent gathered data analysis. The virtual pilot's desk is the first implementation of a jack piping microtunnel- ing machine simulator by means of pilot's steering know-how capture methodology.