In principle, an editor probably should not take one's prior performance into account, either as a reviewer or in any other capacity, and should judge solely on the current work. The workflow of most journals I have dealt with supports this notion, in that there is no way even to readily examine other actions by an author.
In practice, however, a) editors are human, and b) poor performance in one area is often an indicator of poor performance in another related area.
There are, however, two mitigating factors: first, there are often many editors, and second, for most editors you'd have to be mightily annoying to be memorable as a bad reviewer.
Bottom line: yes, if might affect your papers, but you probably shouldn't be particularly paranoid about it. Don't use this as an excuse for doing bad reviews, however: it's far better to withdraw honorably from a review you cannot do well.