Well, whilst it's 'OK' for someone who has just started modelling, there are a lot of problems, and it is breaking most of the cardinal rules of SDS poly modelling :)
1. There are ngons (polys of more than 4 sides), and there is no reason to tolerate a single one of these, which should be considered true modelling mistakes and eliminated / solved to quads. Even if these do not lead to shading errors in the render it is still bad technique to leave them unfixed. Other modellers cannot respect people who leave ngons in their meshes ;)
2. There are triangles, which also don't need to be there. A lot of the art of SDS modelling involves solving all triangles to quads.
3. Complex poles (more than 5 edges converging on a single point) cause shading errors if they are not on perfectly flat surfaces, and even when they are on flat surfaces, are still technically considered bad technique. You have a massive one of those in the middle of your model that should be replaced by a quad patch ideally.
4. Edge flow - this mesh doesn't really have a considered edge flow, so your polys look messy and random and not thought out.
5. Uneven poly sizes / unnecessary edges. This model suffers from that problem too - there are edges where we don't need them, and some wildly variant polygon sizes, which we should aim to minimize in good quality modelling...
6. Uneven circular point distribution - you can't get a flawless cylinder shape under SDS if you don't have nice even segmentation to its surrounding topology. Note the pinching on your forward hole as a perfect example of this...
7. Reference matching. This isn't going very well either - you are missing some of the defining features and curves in your reference pics, most notably the center section and gentle dome of the yellow top cover, and the interesting bulges / ridges in the metal box underneath it.
But all that said, I have ludicrously high modelling standards, and it renders OK, so can't be considered a total failure. Hope that helps.
CBR