Abstract: With the development of precision agriculture technology, agricultural robots which improve efficiency and save labors for agriculture are used for information collection in field, instead of the traditional manual recording and observation. At home and abroad, most of navigation researches of walking platform meet the requirement of tracking precision. However, they mainly use RTK-DGPS (real time kinematics - differential GPS) positioning system, and it is so expensive that it restricts the large-scale promotion of automatic navigation of walking platform to some extent. Consequently, aiming at uniform speed and line tracking problem in complex field environment for information collection platform, a motion controller with transverse correction and longitudinal constant speed control was designed. This paper constructed motion control hardware system using information collection platform as the carrier, STM32F4 32-bit microcontroller was as the control center and multiple sensors were used for obtaining the vehicle status information, including 3D (three-dimensional) electronic compass, low precision Beidou module, rotary encoder and angle sensor. The working principle of the controller could be described as follows: Firstly, 2 points were chosen to determine tracking path. Then, the current position information and the heading angle of the platform in the field were obtained by Beidou module and electronic compass. They were used to calculate the lateral offset and course deviation, which were 2 inputs of fuzzy controller, and the output of fuzzy controller was got through designing control algorithm to control steering angle of front wheel. It ensured information collection platform to always track the goal line. Meanwhile, speed information of the platform was measured by the rotary encoder, which was the input of PID (proportion, integral, derivative) controller. It realized constant speed walking through speed feedback regulation. On the other hand, in order to improve the positioning precision of Beidou module, a method was presented, which was that 3 Beidou modules of low precision were placed in the way of an equilateral triangle, and average positioning data of 3 Beidou modules were obtained at the same time. Then, they were used as the final location data of the center point of the equilateral triangle, which took advantage of compensatory principle of random error to reduce the random error and improve the positioning precision of the Beidou module. By this way, the average positioning precision of the Beidou receiving module was increased from 2.06 to 1.50 m, and the dynamic positioning precision was improved to within 0.78 m. The speed tracking experiment was carried out in the field of Modern Agricultural Science and Technology Experiment Site of Huazhong Agricultural University. The result showed that the control system could work steadily when the vehicle ran at the speed of 0.4 m/s; Proportion coefficient was 100, the differential coefficient was 40 and the integral coefficient was 50. The adjustment time was less than 3 s and the maximum error of speed was less than 7%. When running at the same speed, the initial lateral offset was set to 1.4, 2.0 and 2.5 m, the time required for stable tracking was 11, 15 and 25 s, respectively, and the maximum linear deviation after stabilization was all less than 0.31 m. The research of motion controller on information collection platform not only promotes acquisition ability of field information but also provides technical support for the efficient and intelligent operation of information collection platform in field. At the same time, it offers the possibility of greatly reducing cost in linear tracking, which is of benefit to the promotion of precision agriculture.