I don't really understand what you mean with the question, but still I try to explain.

As you are talking about Illumina

* Almost all Illumina platform produces paired end reads ( except user specification of only single end run)

* Paired end is always better and preferred over single end as it will help in resolving contiguous sequence (this is what illumina calls it)

* Longer reads are always better than shorter reads thus 2x300 Miseq would be better in the bioinformatics pipeline that 2x150 platform.

But before going for the sequencing, brainstorm over the sequencing cost and bioinformatics. As this is the major factor to consider it should be resolved. Since, both platform can serve the purpose and there are various pipeline available for it, you have to choose what suits you and than go for sample prep and sequencing.

I did not understand what you mean with 300 nt INSERT ?

I think going for 150 bp read will be more efficient, especially that you have a reference. Having a 250 bp will result to an overlap of ~ 100 bp. This overlap will simply be a waste in sequencing. By going for 150 bp, you will have be able to get higher depth across your fragment with the same amount of sequecing (i.e., total bases sequenced). And having higher depth is more ideal in your assembly. But again, you already have a reference, and if the reference is good, there might be no need for very high depth as well (e.g., resolving problematic bubbles when performing a graph based de novo assembly).

Thanks Abhijeet Singh and Ambrocio Melvin Agulan Matias for valuable inputs.