Abstract: Abstract: In precision agricultural equipment automation testing process, sensors on agricultural vehicles for collecting information need to keep the relative position with the ground from impact of narrow, uneven and complicated conditions. In the requirements of obtaining accurate attitude and position about the sensors relative to the ground level and to alleviate this problem, a two-axis attitude control system based on MEMS sensors was proposed. Firstly, the operational principle and control methods were analyzed. The Euler angles coordinate transformation method was introduced which is used to describe objects position accurately in inertial space. Secondly, the hardware and software of the attitude control system were designed. The system hardware includes power supply, accelerometer and gyroscope sensors, stepping motor drive, LCD and serial communication. The control system had two axis of rotation called pitch axis and roll axis, which was driven by stepping motor in each axis. And the system was controlled by microcontroller to drive two stepping motors to keep the relative level with the ground. What is more, the accelerometer is used to measure the acceleration of gravity in the condition of static, and the gyroscope is used to measure the angular velocity, so that the attitude control system used these two MEMS sensors to estimate the platform angle. The data of the gyroscope and accelerometer are real-time sampling to gain and analyze the current attitude angle. Therefore, multi-sensor information fusion algorithm and attitude estimation model were established. The complementary filter with two kinds of sensors was used to estimate the angle which had a time constant to define where the boundary between believing in the gyroscope and believing in the accelerometer is. Specifically, for the time periods shorter than time constant, the gyroscope integration takes precedence and the noisy accelerations are filtered out. And for the time periods longer than time constant, the accelerometer average is given more weighting than the gyroscope. Thus the control system chose the appropriate parameters to estimate the accurate angle which meet the requirement of system maximum control error. Finally, The PWM signal was output by microcontroller driving to the stepping motor. Then PWM signal was sent to stepping motor actuator to make the stepping motor turn a step angle. Additionally, the PD controller was used in the system in order to achieve rapid control in order that the two-axis direction was driven by stepping motors which made a quickly response to eliminate the horizontal deviation and kept the relative position with the ground. Two absolute encoders are also fixed on each axis. They were used to keep the system automatically reset to the initial position when the system started to work. The test results show that the system runs stably and the method of estimating angle is suitable for low and middle speed changing conditions. Single axis attitude control accuracy is less than 0.5 ° in the condition of flat slope of the road and is less than 3.0 ° running in the bumpy field road. The stabilized platform system meets the demands of the attitude automatically adjust to the constant orientation in the information collection process and attitude estimation process.