Sri Lanka remains a very popular tourist destination in South Asia and the country’s economy is heavily dependent on the benefits of tourism including earnings from tourism as well as direct and indirect employment. The research aims to estimate tourism demand in Sri Lanka in near term and medium-term time horizons (2024–2027) based on a machine learning (ML) modelling approach. The study considered five ML models, consisting of multiple linear regression (MLR), K nearest neighbors (KNNR), gradient boosting regression (GBR), extreme gradient boosting regression (XGBR) and support vector regression (SVR) ML models. The independent variables considered include monthly tourist arrivals with a 12-month lag, Sri Lanka’s exchange rate and consumer price index as macroeconomic variables and several dummy variables to capture events that resulted in a notable shock to tourist arrivals. Based on model performance, the MLR model gave more practically cohesive forecasts, especially when considering seasonality of tourist arrivals to Sri Lanka. The results suggest almost a doubling of tourist arrivals from current levels by 2027.