An important tool in the field of Permanent Magnet Synchronous Motors (PMSMs) is parameter identification, which allows for final stage inspection, adaptive control and condition monitoring. Common identification approaches require the PMSM to be mounted into a test-bench with encoders, which is complicated or not possible for some machines. In order to address this issue, this work combines a Recursive Least-Squares based parameter identification algorithm with the sensorless control technique Direct Flux Observer (DFO). The developed identification technique is implemented on a microcontroller and experimental results on different kind of PMSMs are shown and discussed.