Abstract: This research focused on improving performance of unmanned farm combine harvesters in complex terrain, especially precision and response speed of cutting table height control．A height detection compensation method based on double inclination sensors was proposed．Correlation model between inclination angle and height of cutting table was established by fitting least square method, and correlation coefficient was 0.9958, which significantly improved measurement accuracy．Then, PID control strategy optimized by swarm intelligence algorithm was introduced, Bernoulli chaotic mapping, particle swarm optimization（PSO）and T-distribution perturbation were used to improve dung beetle optimization algorithm, which solved problem of low precision and slow response of traditional PID control．Based on these optimization, IDBO-PID controller has better response speed and stability than DPO-PID and traditional PID in simulation comparison．Experimental results showed that improved control strategy could make rise speed of cutting table reach 0.44 m/s, fall speed reach 0.32 m/s, and height error was controlled within 0.02 m, which could meet operation requirements．