This paper is concerned on finding the best rainfall simulation model for the hourly series that could describe accurately various stochastic properties of the underlying rainfall processes. Two stochastic rainfall-modeling schemes were assessed for their accuracy and suitability in simulating the hourly rainfall series. The first model is a cluster point process model where the clustering feature of the rainfall processes is following the Neyman-Scott and is referred as the Neyman-Scott Rectangular Pulse (NSRP) model. This model was improved by combining the use of transition probabilities instead of autocorrelations in the fitting procedures and Mixed-Exponentila distribution to represent the rain cell intensity. The second model represents a combination of the first order two-state Markov Chain for describing the rainfall occurrence process and the mixed-exponential distribution to represent the distribution of rainfall amount on wet hours. The models were assessed on a 10-year hourly data at KM 16 Gombak, Kuala Lumpur. Box-plots and Root-Mean-Square error was used to compare between the two models. Results of these assessments indicated that both models were able to adequately describe the seasonal variation of many statistical properties of the rainfall processes for various time-scales. However, the NSRP was found to be more accurate than the MCME in the estimation of the hourly rainfall characteristics. MCME with the daily data was able to estimate the daily rainfall characteristics better than using NSRP at 24h aggregation.