The availability of influent wastewater time series is crucial for assessing the performance of a wastewater treatment plant (WWTP) under dynamic flow and loading conditions. Given the difficulty of collecting sufficient data, synthetic generation may be the only option. Usually, the main constituents of the influent time series (e.g. flow, COD, TSS, TKN) show periodic, auto-correlation, and cross-correlation structures in time. Therefore researchers have used statistical models (e.g. auto-regressive time series models) for random generation of the influent time series. However, these regular patterns in time could be significantly distorted during rain events (wet weather flow (WWF) conditions) in which the amount and frequency of rainfall affects the flow and other constituents of the influent. To tackle this problem, a hybrid of statistical and conceptual modeling techniques was adopted. The time series of rainfall and influent in DWF conditions (i.e. inputs to the conceptual model) were generated using two types of statistical models (a periodic-multivariate time series model for influent in DWF conditions and a two-state Markov chain-exponential model for rainfall). These two time series serve as inputs to a conceptual model for generation of influent time series during WWF conditions. The effect of total model uncertainty on the generated outputs was also taken into account through a Bayesian calibration and communicated to the user by constructing uncertainty bands with a desired level of confidence. The proposed influent generator is a powerful tool for realistic generation of the influent time series and is well-suited for probabilistic design of WWTPs as it considers both the effect of input variability (i.e. time variation in rainfall and influent composition during DWF) and total model uncertainty in the generation of the influent.