I don't believe you need either. On the import step, it lets you decide what "destination" field you want the data to go in, and the current field name does not have to match. The order doesn't seem to matter since you designate each field with a destination field (so say your CSV has FName, you choose "First Name" as the destination).
To make it easy on myself, I loaded a 'dummy' record and populated every field, then exported that in a CSV format. That gave me the field names and orders of what MM has as data. I then just tweaked the field names on my import file to match the exported file, and set the fields in the same order that MM had on the dummy file. This made it much smoother to import since visually I had a 'one to one' match during the import process.
One tricky part are dates...they must be formatted in a specific way (yyyymmdd or yyyy-mm-dd if I remember correctly), so if you are using Excel to generate your CSV, you have to remember to fix the date format every time you save out to CSV. Opening a CSV in Excel to review may switch that format around, so if you modify it, you have to again go back and fix the formatting.
