Reactor neutrinos have been an important tool for both discovery and precision measurement in the history of neutrino studies. Since the first generation of reactor neutrino experiments in the 1950s, the detector technology has been greatly advanced.New ideas, new knowledge, and modern software also enhanced the power of the experiments. The current reactor neutrino experiments, Daya Bay, Double Chooz, and RENO have led neutrino physics into the precision era. In this article, we will review these developments and accumulations, address the key issues in designing a stateof-art reactor neutrino experiment, and explain how the challenging requirements of determining the neutrino mass hierarchy with the next generation experiment JUNO could be realized in the near future.The first proposal to detect the free neutrino was to use nuclear bombs. Then it was superceded by the approach of using a fission reactor [12] with the technology of liquid scintillator in the 1950s [13]. The Hanford experiment [14] made the first attempt to search for the neutrino. Shielded by paraffine and lead, its 300 liter liquid scintillator was viewed by 90 2-inch PMTs. No neutrino signal was found due to high backgrounds. The detectors were moved back to Los Alamos and put underground, which confirmed the backgrounds were from cosmic rays. The lesson was clear: "it is easy to shield out the noise men make, but impossible to shut out the cosmos". Therefore, the detector has to be put underground.Besides the passive shielding of the cosmic rays, the first anti-coincidence detector was developed in the later Savannah River experiment to distinguish backgrounds. The antineutrino signature was a coincidence between the prompt e + annihilation signal and the microseconds delayed neutron capture on cadmium. The detector was deployed close (∼ 11 m) to the Savannah River reactor and 12 m underground in a massive building. Finally the existence of neutrinos were convinced [1].Immediately after the observation of neutrinos, Pontecorvo [15,16], Maki, Nakagawa and Sakata [17] suggested that the neutrino may oscillate from one flavor to another as it travels, which stimulated searches for neutrino oscillations. The first experiment on neutrino oscillations was performed in 1979 near Savannah River [18]. It was the first experiment that use heavy water to detect neutrinos [19], by neutral-current interactions (NC: ν e + d → n + p + ν e ) and charged-current interactions (CC: ν e + d → n + n + e + ). The neutrino oscillation signal, expressed as the double ratio of the measured to theoretical rate of the CC with respect to NC process, was measured to be 0.40±0.22, a first indication of neutrino instability. It was interpreted as neutrino oscillation and deduced an allowed region of ∆m 2 and sin 2 2θ assuming two base mass states to be involved.A long debate followed about the indication of neutrino oscillation. The ILL (Institut Laue-Langevin) experiment [20,21] in France used 377 liter LS placed at 8.75 m from the ILL reactor (57 MW, 93% 235 U). For the ...