||Nowadays, a series of emerging ultra low power applications (e.g. picoradio, smart dust, implantable medical microsystems, wireless sensor network) that demand compact, low cost, long lifetime and high integration, have potential to be used in many areas, where ubiquitous computing, sensing, and perception facilitate the interaction between human and the environment. In these applications, providing the required supply voltage and power is a big challenge. The conventional solution is to use electrochemical batteries. However, battery has limited energy capacity, relatively large volume with respect to the electronic circuit, finite recharging cycles and is difficult to be replaced regularly in many cases. All the above disadvantages pose a big limitation on the wide deployment of such systems. Hence, eliminating the batteries is much desirable for these applications. On the other hand, the average power consumption for most ultra low power applications can be down to the level of hundreds or even tens of microwatts. In such low power level, power scavenged from the environment can be used as an alternative power source to provide a virtually infinite lifetime. Mechanical energy conversion by piezoelectric materials is one of the feasible approaches for energy harvesting. In this thesis, the design of a mechanical vibration energy scavenging and power management system is proposed for ultra low power applications. A new maximum power point tracking scheme is proposed for piezoelectric conversion to achieve the highest energy harvesting efficiency. This scheme consumes very little power and features simple hardware implementation, so it is especially suitable for ultra low power energy harvesting applications. This proposed system is capable of self-starting and self-powered operations, thus eliminates external battery integration and significantly reduces the system volume and cost. In addition, a micropower DC/DC buck converter is designed for ultra low power applications to achieve higher power conversion efficiency. System modeling, analysis, and VLSI implementation were developed in this thesis. A hardware prototype chip for the proposed system was fabricated and tested. Simulation and experimental results show that the proposed MPPT scheme can track well with the vibration status changes and achieve a higher than 90% energy harvesting efficiency.