If I understand well, you are talking about to use the function `predict()`. So you have the model: `LM <- lm(PSR ~ Area+Forests, data = Wetlands)` The function `predict()` provide this usage `pred <- pred(your_model, your_data_test)` but `your_data_test` have more variables than your model, right? You can make a `new_data_test` using this, `new_data_test <- data.frame(your_data_test$variable1,your_data_test$variable2)` `pred <- pred(yourmodel, new_data_test)` I suppose that this may be works well.