A self-powered system for the Internet of Things (IoT) is demonstrated for efficient energy harvesting of naturally available mechanical energy. In this system, new contact-separation mode triboelectric nanogenerators (TENGs), based on fluorinated ethylene propylene, are investigated using the segmented multi-TENG configuration to reduce the effect of parasitic capacitance. The TENG extraction is optimized using a unit step excitation involved with the Dawson function to achieve a high voltage (400 V) and a high current (26.6 µA). To fully extract the power of the TENGs, the power management integrated circuit (PMIC) specially designed for adaptively controlled, high-voltage (HV) maximum power point tracking (MPPT) is proposed. The PMIC implemented in a bipolar CMOS-DMOS 180 nm process can handle a wide input range (5-70 V) by consuming 420 nW. The MPPT control allows a wide range of impedance matching from 10 to 300 MΩ, achieving a tracking efficiency of up to 98.2%. The end-to-end efficiency of 88% demonstrates state-of-the-art performance. To supply a higher instantaneous power than that available from the TENGs, a duty-cycling technique is successfully demonstrated. The proposed energy harvesting system provides a promising approach to realizing sustainable and autonomous energy sources for various IoT applications.