Abstract: A statistic analysis was conducted for surface relative elevation data from typical field experiment under different survey grid spaces. The feasible survey grid space for the field with small dimension was discussed by comparison of the manpower input of the survey, description of the microtopography conditions and the precision of the estimated earthworks. The results show that the precision of the estimated earthworks, manpower input of the survey and description of the microtopography conditions are vital factors for determination of the feasible grid space. A 10～15 m grid space is suggested for land levelling survey in order to achieve the suitable precision of the estimated earthworks and less manpower input. The land levelling design for the field with small dimension could be based on such data from suggested grid space with the benefits of the higher precision and lower cost.