This paper reviews the design techniques of the electrical neural recording system for energy-efficient operation. To operate neuroprosthetics by decoding action potentials (or neural spikes) and local field potentials in vivo, the electrical neural recording system can be used. For extending the system battery lifetime in vivo and avoiding tissue damage due to the heat, the neural recording system needs to be energy-efficiently driven while consuming low power and achieving low noise performance. In this paper, three energy-efficient design techniques are reviewed with basic properties of the electrical neural recording system: 1) energy-efficient input structure of the operational transconductance amplifier, 2) neural recording front-end channel using the low supply voltage, and 3) neuronal activity-dependent recording system.