Thanks to the sensor miniaturization and efficient edge computing, the maintenance of mechanical equipment can now successfully evolved from conventional run-to-failure and time-based preventative maintenance to condition-based maintenance or even predictive maintenance. This article takes the intelligent vibration sensor developed by ITRI as an example to interpret key issues such as hardware selection and integration, signal processing, and system verification. It is expected that the readers can therefore get an in-depth understanding of how the integrated sensing, analysis, and diagnosis solution can help on improving the efficiency of machinery prognosis and health management.