Dissertation (MSc. Statistics)Tourism activities are beneficial worldwide both socially and commercially and Tanzania is not exceptional. The sustainability of these benefits depends on the continuity increase in the number of tourists. This increasing flow of tourists is the product of tourist’s satisfaction of which among other factors is a pre-plan of activities. This study aimed at forecasting of Tourist Arrivals at Serengeti National Park in Tanzania using the time series analysis. The study adopted a longitudinal research design, with the monthly secondary data from January 2001 to December 2017 obtained from the Tanzania National Park Agency (TANAPA). The forecasting process was led up by analyzing data which showed indicators of non-stationary. The findings indicate that both local and foreign tourists increase over years at the SENAPA. The Autoregressive integrated moving average (ARIMA) models was adopted in answering objectives. The Akaike’s information criterion (AIC) and the Bayesian Information Criterion (BIC) was involved in the selection of the best forecasting models for tourists’ arrivals. It was found that the ARIMA (0, 1, 2), ARIMA (2, 1, 2), and ARIMA (2, 1, 1) were a fit for domestic, foreign, and the combined tourist arrivals data respectively. The findings revealed an increasing positive trend which means that in the forecasted period of 24 months from January 2018 to December 2019, the number of tourists will rise. The study recommends the use of other different models in order to make comparison of model results.