question on multiple linear regression

hi all, i have a rough idea of how multiple linear regression works(using regression coefficient and error term to predict a single y) but i'm not sure how does it make a series of prediction for different values of y with different values of x ?

example i have 300 values for each of the 5 variables x and 300 values for y. I used 250 for training and the rest for testing. so how do multiple linear regression makes the prediction ? Does it average the 250 equations ?

[dpb]

No, you evaluate the fitted model at values of the independent variables. Those could be the values in the existing data set or others.