Precision agriculture analyzes the spatial variability according to the characteristics of an optimum setting of agricultural materials. To raise the profitability of agriculture and to reduce the environmental impact, technological research and development of precision agriculture has been conducted. In Asian countries such as Japan and Korea, yield monitoring system and crop growth sensors for rice and dry-land crops have been studied. Market for full feed type combine harvesters is recently growing in many countries including Republic of Korea. Yield monitoring system is one of the recent trends of the combines, and the major components consist of a positioning system, grain flow and water content sensors, ground speed and cutting width sensors. Objective of the paper was to design and construct an ultrasonic cutting width sensor for full-feed type mid-sized multi-purpose combines, as a part of a yield monitoring system, targeting row-planted and broadcasted rice, barley, wheat, soybean, and rapeseed. The target combine harvester was 55-kW full feed types for various crops with about 20% of the grain water content. Cutting width was about 200 cm, a maximum working speed of 1.7 m/s, and the turning radius of 1.5 m, and overall loss of 1.5% or less. First, an experimental device with paired ultrasonic sensors was designed and constructed. Two ultrasonic sensors (UDS-10A) were mounted on a frame, shaped of a combine header, connected to a computer through an USB port, and the data were obtained with customized software. Then, calibration tests were conducted in various conditions. Basic performance of the ultrasonic unit was validated by confirming distances to the wall. The tests were conducted from 50 to 200 cm with a 10-cm interval, and the results proved the accuracy, showing that the coefficient of determination of the linear regression was 0.99. Average and maximum errors of the tests were 1.7021 cm and 0.9686 cm, respectively. The experimental unit was also tested for field conditions. Distances from the crop were varied from 0 to the full cutting width (i.e., 210 cm), and the signal was collected for 1 minutes at stationary condition with 3 replications. Average and maximum errors were 2.5195 cm and 4.8021 cm, respectively. Next, in order to know the possibility of the measurement of actual combine width with the above experiment method, the distance with the crop has been moved (around 1.7m/s) for the measurement. Average and maximum errors were 2.2934 cm and 5.4763 cm, respectively. However, the value was not adequate for the measurement of cutting width, which was minimum 82.3% which must have used all the data. Results of the study showed a good potential of the fabricated cutting width sensor. Future study would include dynamic tests, combine installation and field tests during harvesting season, and optimization with other components of the yield monitoring system. Also, future study will develop a data process algorithm for the precise real time measurement of cutting width.