This study was carried out to develop a system and methodology to simulate the engine load variation occuring during agricultural field operations for a laboratory engine test. The system consisted of an electric dynamometer, an Apple II microcomputer, and a data acquisition and control system. Several pieces of instruments were utilized to measure various engine performance data. Both engine torque and engine speed were fully controlled by a computer program. The dynamic characteristics of the system were analyzed through a series of tests and the limitations on the load simulation test were presented. The results of the study are summarized as follows: 1. Engine speed and toque were controlled by a computer program. The use of a stepping motor and reduction gears enabled engine speed be controlled within 1 rpm. 2. The natural frequency of the dynamometer-engine system was found to be around 5 Hz, at which the load simulation would be impossible because of resonance. 3. For the harmonic inputs with the frequencies above the natural frequency, the signal attenuated too much and therefore the load simulation was impossible. 4. The step response of the system showed an overshoot of 24.5 percent and the settling time for 5 percent criterion was around 3 seconds. 5. When actual field test data are utilized for load simulation, a low-pass filter should be included to attenuate the frequency components around and above the natural frequency.