With the development of autonomous driving technology, lidar technology has also been rapidly developed. As one of the Lidar systems, Frequency Modulated Continuous Wave (FMCW) calculates the distance of the target by detecting the beat signal frequency of the local light and the signal light. It can also calculate the velocity of the target by the optical Doppler effect. Compared with traditional time-of-flight lidar, FMCW lidar has many advantages. First, FMCW utilizes the principle of coherent detection to filter out the background noise effectively. The peak power of the continuous-wave is small, which is safe for human eyes. Secondly, the FMCW system can detect both the speed and distance of the target. Thirdly, when it comes to long-range detection, the FMCW system doesn't need a Single photon detector, so it can reduce the cost. However, in the field of autonomous driving, with the increase of imaging speed, the laser modulation rate has to increase, and then the nonlinearity of the laser source rises. as a result, the frequency of the beat signal varies greatly and the precision and resolution of the detection become worse. The resampling method, photoelectric negative feedback method, and pre-correction method can solve this problem. Considering the cost of the FMCW system, we use the pre- previous studies based on the auto-driving application. This paper studies the nonlinear pre-correction of DFB laser in high-speed tuning by simulation and experiments at the same time. Here, we first briefly introduce the principle of FMCW detection and analyze the influence of optical nonlinearity on detection precision. Secondly, we carry out the DFB simulation and use the Simulink toolbox of MATLAB to establish the temperature dominated tuning model of the DFB laser. Then, we assume the parameters of the DFB laser and analyze the nonlinearity distortion with the modulation rate of 100 kHz. Because the temperature hysteresis has most of the influence on optical nonlinearity, here, we do not consider the carriers' effects on frequency turning. After that, we use the iterative algorithm to correct the nonlinearity of optical frequency for many times and analyze the convergence of this algorithm. The correctness of the DFB module and iterative algorithm is verified. There are three steps in the iterative algorithm. At first, we calculate the optical frequency by STFT or Hilbert transform from the beat signal, then we calculate the differential value between two adjacent time points in the arbitrary wave generator and figure out the relative value of each point. At last, after the linear transform, we can get the real value of arbitrary wave generator at each time. Based on the simulation, we first set up an experimental platform to correct the nonlinearity of the existing DFB laser with two times of iterations. To analyze the effect of the nonlinear correction method, we detect two adjacent targets with the distance of 51.5 m and 60.5 m. The correctness of the simulation model and an iterative algorithm is verified. The results of the simulation and experiment show that the iterative method can effectively reduce the nonlinearity of the light source. With the increase of the number of iteration times, the nonlinearity improvement of the light source becomes more obvious. The nonlinearity of the lower and upper area of the optical frequency of the light source decreases from 0.050 0 and 0.020 0 to 0.004 2 and 0.002 6 respectively after two times iterations, which is consistent with the simulation. The detection results of two adjacent targets show that the 3 dB bandwidth of the detection system is effectively reduced after correction, which proves the feasibility of the iterative algorithm and modulation module.