North Sumatra is one of the areas in Indonesia which has various tourist attractions with its natural beauty. The potential of North Sumatra makes the attraction of foreign tourists want to visit the existing tourist attractions in the area. Every year foreign tourists who come to North Sumatra, especially tourists who come from the ASEAN region the number is changing. Therefore an estimate is needed to determine the number of tourists who come and later this estimate is useful for the government to anticipate the surge of tourists who come. The data to be estimated is the data of the number of tourists who come to North Sumatra according to national origin sourced from Central Bureau of Statistics of North Sumatra (BPS Sumut) in 2011-2015. The algorithm used to make estimates is the artificial neural network Backpropagation. There are five architectural models used in this estimate that is, 3-7-1 has 100% accuracy rate, 3-10-1 = 62.5%, 3-11-1 = 75%, 3-12-1 = 88% , and 3-15-1 = 87.5%. The best architecture of the five models is 3-7-1 with 100% accuracy rate and MSE of 0,006513Keywords—Estimates, Foreign Tourists, North Sumatra, Neural NetworksDOI : 10.5281/zenodo.1402836