Tourism revenues are a significant source of income under the current account service item of countries. These revenues are not included in exports, despite being compared with the export revenues of the countries and in economics the ratio of tourism revenues to export revenues is used as an indicator. In developing economies, tourism revenues play a role in closing the current account deficit. The prediction of this rate in countries with foreign trade deficit is important in developing tourism, export and import policies for the future. In this study, multiple linear regression method (MLR), one of the traditional methods, and the artificial neural network method (ANN), one of the machine learning methods were used to estimate the rate of tourism revenues of the sample country Turkey to its export revenues. In the model of the study covering 2004-2020 period, the number of tourists received, total income from tourism, average expenditure by tourists per capita, population, total export revenue, growth rate, Euro/TL and US Dollar/TL rates were chosen as independent variables. As a result of the study, the R2 value was found to be 91.7% for ANN and 90.8% for MLR which were very close to the ideal value. According to the predicts made on the model developed based on this, the rate of Turkey's tourism income to total export income in 2025 is estimated as 31.83% according to ANN; 32.73% according to MLR while in 2030, it is estimated to be 33.25% according to ANN and 36.78% according to MLR.