Dissertation (MSc Statistics)This study aimed at modelling and forecasting of the daily stock prices of NMB and CRDB on a short-term basis by using the autoregressive integrated moving average (ARIMA) models and exponential smoothing models. It was guided by three specific objectives, which were: to fit models to the daily stock price data of NMB and CRDB banks in Tanzania, to select the best fitted model of CRDB and NMB by using model selection criteria and; to forecast and evaluate the forecasted daily stock prices of CRDB and NMB banks in Tanzania by the best-fitted model. This study was conducted in Dar Es Salaam Region, Tanzania. The modelling process was preceded by analysing the time series of interest which revealed the presence of non-stationarity. The results indicated that, On fitting the model, the resultant models were: ARIMA(3,1,1),SES,DES and DTLES for NMB whose their parameters were statistically significant while for CRDB the resultant models were ARIMA(1,1,2), SES and DES whose their parameters were statistically significant for. Among the fitted models, the following were selected through Akaike’s information criterion (AIC) and Bayesian Information Criterion (BIC): for NMB the DTLES was selected as best model while for CRDB, ARIMA(1,1,2) was selected as best model. Stock forecasting was based on the best selected models which gave the forecast daily stock prices of NMB and CRDB banks in Tanzania. Also, it was revealed that, the developed models fit well in the historical data and can be used in short-term forecasting of the future values of NMB and CRDB bank in Tanzania. However, for CRDB, some forecast did not agree strongly with the observed values. Therefore, it is recommended for future researchers to consider non-linear models such as the ARCH model together with its variants, probability distribution fitting and SVM models in addition to the ARIMA model.