You didn't specify whether you have one Y and many X's, or many Y's and one X. I will assume the former.

If you want to analyze the Y separately against each X, then use Fit Y by X. Put the Y variable in the **Y, Response** role, and put the other variables in the **X, Factor** role. This produces a separate model for each X.

If you want to analyze the Y against all the X's at once in a joint model, then use Fit Model. Put the Y variable in the **Y** role, and make the X's as Model Effects using the **Add** button. This produces a model which predicts Y as a function of all the X's at the same time.

Without knowing more about your situation, I can't say more about which way is best.