Hello Derek,
I am not familiar with SAS, but your first merge between household and individual women's data looks correct. The result of the merge should be all individual women with household information added to them (12674) and some cases with only household information and without individual woman data, because no women were interviewed in these households. The identification variables are indeed V001 (cluster), V002 (household) and V003 (woman's line number)
You then went on to merge this file with the male data file. Here the identification variables are MV001 (cluster), MV002 (household) and MV003 (man's line number) renamed to V001, V002 and V003. Since the line-numbers of all household members in the same household are unique, V003 and MV003 will never be the same. I am not sure what happens when you merge the man's data to the already merged household/woman's file, but believe that the man's data will simply be added to the household/woman's file.
Depending on your desired unit of analysis: if you want all women and men as unit of analysis with the household information added to them, then I would first add the women and man's data together (10826+4121 = 14947 cases) and then merge this with the household data. In this case you will get a file with all interviewed women and men as separate cases with the household information added. If you want to create couples files then variable V034 needs to be used (linked to MV003) and then the household data could be added.
It would be more difficult to create and analyze household information, with all data for interviewed women and men in that household added to the household information (in this case you would have 10826 cases).
Let us know if this is not clear or does not answer your questions.