In order to achieve the maize sowing time decision-making and improve the effective accumulated temperature of maize growth, it is needed to understand the soil spatial variability characteristics. Totally 20 sets of wireless sensor network nodes were deployed in Zhaoguang Farm in Heilongjiang Province for one month in 2015. In addition, two handheld mobile sensor nodes were chosen to increase the measurement number. According to the method, soil temperature and moisture data were obtained from 5 th to 29 th in April with 240 m×240 m, 120 m×120 m, 60 m×60 m and 30 m×30 m grids. The isotropic and anisotropic variation characteristics and distribution patterns of soil temperature and moisture were analyzed based on statistics semivariance function theory and GIS space Kriging interpolation method. Experimental results showed that the semivariance of soil temperature and moisture were suitable for the spherical model and exponential model, respectively. Both of them had strong spatial autocorrelation. The distribution of soil temperature was block with autocorrelation distance of 51.56 m. And the distribution of soil moisture was ribbon with autocorrelation distance of 154.16 m. The anisotropy of soil temperature and moisture variation was also significant. The soil temperature variations in 45° and 90° directions were significantly greater than those in 0° and 135° directions. The soil moisture determination coefficient ( R 2 ) was 0.77 with significant variations in 0° and 135° directions. The research results provided a scientific guidance for the decision-making of maize seeding time and the determination of soil sampling distance.