One way to do this is to fit a linear regression model to pooled data (data set 1 + data set 2), with intercept and slope varying across two data sets. Then the test for equality of intercept can be done easily. If you still don't get it, I guess many experts here can answer your question.