The correct weight for INCWAGE is ASECWT. This is assuming you are interested in making time series of occupational wages, and not linking individuals over time. If you do want to link individuals, I recommend perusing some of the IPUMS training materials on linking the CPS.
As for calculating average incomes, whether the top-coding affects the mean depends on the year and the subsample you are interested in. I recommend reading the documentation on the income topcodes and swap values. If you’re looking to calculate mean incomes by occupation, the topcodes will affect your results. However, depending on the occupation, this might not matter much. It should only significantly affect very high wage occupations. The rank proximity swap technique (available starting in 2011) should minimize the effects of the topcodes, even within subgroups like occupation.
I also recommend checking out the Occupational Employment Statistics program at BLS for another source of data on this topic.