Multivariate analysis was used to build macroinvertebrate predictive models for stream assessment in Britain, Australia, and the west coast of United States. The philosophy behind these predictive models is similar, but variations exist and are adapted for different regions. A macroinvertebrate predictive model in Maryland has been improved using different Region–style methods, including the Assessment by Nearest Neighbour Analysis (ANNA) and Region of Influence (ROI) methods. For better prediction precision, different parameter selection schemes (stepwise AIC, exhaustive AIC, and exhaustive BIC, where AIC is Akaike's information criterion and BIC is Bayesian information criterion.) and rational multiple regression function checking has been used to prevent overfitting of the model. Root mean squared error (RMSE) was used to select the final best or better performance model. The calibration results from Region-Style macroinvertebrate predictive models are better than previous built RIVPAC-style model. The different parameter selection criteria discourage overfitting and improved the prediction results in validation data. Region–style methods can be alternative methods for building predictive model.  The different selection schemes offer better parameter selection strategy. The best performing ANNA and different ROI predictive models can later on be automated in the GIS environment.