The Sacramento-San Joaquin River Delta (simply the Delta) constitutes the heart of California's water supply and economy. The growing importance of the Delta has led to the development of several numerical models to predict hydrodynamic as well as water quality conditions. The past few years have seen major changes in model development with the advancement from one- and two-dimensional (1- and 2-D) models to 3-D models. These hydrodynamic and transport models will play, in the future, an important role in the management of the Delta, with regard to climate change (for instance sea level rise), water allocations and land use. This project's main objective is to provide an assessment of the most important models, in terms of their ability to predict, with accuracy, flow and salinity in the Delta. The models analyzed are RMA2 and RMA-WAM, both models developed by Resource Management Associates (RMA), and DSM2, developed by the California Department of Water Resources (DWR). For each model, simulations were undertaken using similar input and boundary conditions. The model performance was assessed for several scenarios comprising of relatively wet and dry years and high and low pumping, by comparing model results to observed data which have been collected by various water agencies. Thirty-two stations within the Delta were chosen for this purpose. Plot comparisons and statistical metrics help in understanding the models in depth, and their complexities. Several shortcomings on the observed data and uncertainties in model implementation (due to the complexity of the Delta system) are also highlighted. Plot comparisons include daily averaged data of stage, flow and salinity, displayed over the water year; tidally averaged data; 15-minute/1-hour data, displayed over the month of March; 15-minute/1-hour data, displayed from March to July; scatter plots with their respective R2 values, showing predicted versus observed values. Metrics used in this work are: Phase Errors, Amplitude Errors, Percentage Root Mean Square Differences (RMSD) and Goodness of fit tests (Scatter Plots and R2 values). Our work shows that DSM2 and RMA2 present similar, good prediction levels, with overall RMSD of about 20-25% with respect to field data, for salinity. The predictions of stage varies from very good to excellent with average phase differences less than 30 min, and the prediction of flow discharge is very good in general. The project thus advances a complete and unprecedented characterization of the accuracy of the prediction of models in terms of year and region in the Delta.