In your question, you mentioned that you "wrote the full paper and then realized that it is possible that the particular journal requires experimental validation of any modeling results."
Thus, before answering your questions, I think it will be a good idea to look at what the IEEE says regarding the Letter format. [Also, since you commented on the possibility of submitting to the IEEE Journal of Microelectromechanical Systems (JMEMS) in particular, I will incorporate that venue in my answer below.]
From the IEEE JMEMS site, we find that:
The Letters format is meant to encourage rapid publication of original and significant contributions that are limited in length (three or fewer journal pages), but appropriate for early dissemination.
Further, from the same IEEE JMEMS site above, we find that those seeking to submit their Letter are required to adhere to the following:
All submissions to JMEMS Letters are required to include an accompanying submission letter that puts forth persuasive arguments supporting the need for rapid publication. Possible supporting statements for these arguments might be built using: 1) references to recent literature that illustrate how the Letter addresses an already existing and pressing problem, and/or 2) "impact" statements outlining timely ramifications of the work, and/or 3) clear arguments supporting the case that the Letter points the way toward significant and meaningful changes in previous technologies, materials, or designs.
Note that nowhere in the information from the IEEE JMEMS site above do we find that the Letter format is suitable for those authors who (almost) have a full paper worth of research findings, but are missing key pieces of expected content.
So, let's get to your questions:
No, the Letters option should not be viewed as an easier path to publication. (Surely, if it were easier to publish in the Letters format, then everyone would do that.) In fact, as noted above at the IEEE JMEMS site, authors submitting Letter manuscripts are expected to justify why their work should be rapidly disseminated, and this required justification is on top of having a manuscript which makes a coherent case for the research findings presented.
If you were to submit your manuscript as a full paper, and it was rejected, say, because it doesn't have a sufficient amount of experimental validation, I doubt that submitting it for consideration as a Letter would be successful. Why? In the particular case of the IEEE JMEMS venue, both full papers and Letters are handled by the journal (i.e., both full-length papers and Letters are published in the same venue). Thus, we have to assume that the Editor-in-Chief (EiC) for JMEMS would be aware that your full-length manuscript was rejected on the grounds that your paper was missing key pieces of expected content. Surely, any EiC at any journal of reasonable quality would view this course of action as an attempt to "game the system," and they would thus frown on this strategy rather heavily.
In conclusion, I think the best course of action is to work on the experimental validation of your modeling results, or find another venue where such validation is not seen as a requirement for publication.