Flow measurement in open channels using mobile or portable flumes is an effective solution for accurate discharge predictions. Researchers have developed various models using dimensional analysis, primarily considering upstream flow depth (y1) and flume geometric parameters. The present study aims to improved discharge prediction models by incorporating the effect of critical depth (yc) and ensuring applicability to both conical and cylindrical baffles. By studying flow through a Conical Central Baffle Flume (CBF) in trapezoidal channels, the research utilizes experiments and CFD-based simulations to gather data from six Conical CBFs with varying discharges. Applying Buckingham's it-method, two new discharge prediction models have been developed and calibrated. These discharge models include upstream flow depth and flume's geometric parameters as key influencing variables, while the effect of critical depth is incorporated through these key influencing variables (i.e. y1, B, D, and c). The first model showed an absolute mean relative discharge error of 1.84 %, while the second model exhibited absolute mean relative errors of 1.99 %. Given the geometric similarity between conical and cylindrical baffles, these models were also validated for their use with Cylindrical CBFs, to confirm their broader applicability. Both the developed models were evaluated using statistical indices (RMSE, RME, PBIAS, and NSE), and as the first model is found to be more accurate, it has been proposed to estimate flow rate. The comparison of the developed models with existing models from the literature for both Conical and Cylindrical CBFs demonstrated that the proposed discharge model (Discharge Model-1) exhibited lower mean relative error in predicting flow rates. The study concludes that incorporating the effect of critical depth in development of models improves discharge prediction accuracy, making these models valuable for field engineers using both Conical and Cylindrical CBFs in trapezoidal channels with side slopes ranging from 0.5 to 1.5, for flow rate measurements.