You are correct that some people reflexively dismiss as “junk” any poll whose results they don’t like, which is obviously a stupid thing to do, since failing to face the truth often leads to defeat. But in this case, this is a poll by Suffolk University (a liberal school, and not exactly Mason-Dixon or Rasmussen in the world of polling) of registered voters (not of likely voters, which would have presented a truer snapshot of the electorate by excluding those unlikely to vote) and with a sample that was 41% Democrat, 37% Republican and 23% other (when Republicans have outnumbered Democrats in recent elections in FL). So there are very good reasons to suspect that this poll’s results were skewed in favor of Obama, and that Romney is actually several points ahead of Obama in FL.
As for the fact that the election is a little less than 6 months away so anything can happen, that is absolutely correct, but it’s still important to know where the electorate stands today.
Historically, these Suffolk polls may well be weak, but this particular poll is not an outlier...
http://www.realclearpolitics.com/epolls/2012/president/fl/florida_romney_vs_obama-1883.html
Rasmussen has Florida Romney +1 as opposed to Suffolk with Obama +1. Basically the same result... close race within the margin of error.