We reveal the power low distribution and spatiotemporal variability of the multiscale (1-1000 km) SST macroturbulence statistical characters from the Level-2+ (L2P) satellite orbit products of VIIRS radiometer and high-resolution numerical model over the Kuroshio Extension (KE) region (145°E-160°W, 20°N-50°N), with the methods of structure function and spatial variance. The power low distributions are different between the model and observation especially at submesoscale. The observed power low distribution slope is close to 2/3 throughout the 1-1000 km spatial scales, which is equivalent to the -5/3 wavenumber spectrum. The model-derived slope decreases from ~1.5 at scale r = 1 km  to ~0.5 at scale r = 1000 km. The model results obviously overestimate the SST spatial variance at mesoscale and most part of submesoscale. However, the model did a good job of predicting the phase of seasonal variability. At large-scale, the SST spatial variance show a maximum in winter. The maximum gradually shifts towards spring as the decrease of the scales, and a downscale process is believed occurring during this period. The latitude-dependent spatial variance shows that the strongest SST macroturbulence occurs around 40°N, which is consistent with the latitude of climatological background SST front. It implies that the background SST front is the main source of the strong SST macroturbulence structures. New high-resolution observations are needed to further validate these results, and this work will benefit the oceanic submesoscale numerical model improvement.