Ok, so
We could say that the graph that shows the scatterplot with the linear model that best predicts data is the option D. That's because the points take values around the line and they are close to it. In the other options, the points are not close to the line, and they are not around it.