A frequency modulated continuous wave (FMCW) light detection and ranging (LiDAR) system based on reservoir computing (RC) is proposed and its reliability is verified through simulations and experiments. Fundamentally different from previous approaches where Fourier analysis is always required to determine distance information, here temporal signal analysis using RC is adopted. The system is robust to nonlinearity in frequency modulation of optical carrier, hence improving the range detection resolution. An intensity-modulated light is injected into a semiconductor laser to deliberately generate a non-ideal nonlinear wavelength scanning light source. The change in distance corresponds to the change in the IF signal over time. The RC classifies the IF signal waveform in the time domain to eliminate the need for linearization, and determines the distance from the classification results. The distance measurement resolution of this method is 1 cm in the 6.1 GHz scanning range, which shows that the method effectively addresses the effect of source nonlinearity due to the scanning wavelength while reducing computational complexity. The method has been demonstrated to reduce calculation cost.