To deal with the separatists alone which are not backed by a numerically massive US attack, the solution can be done using 3 different tiers of strike options.
The lowest and most numerous tier will consist of suicide drones, long range MLRS and PGMs of the air force. These will knock out major dual use infrastructure as well as hit force concentration targets of opportunity.
The second tier of weapons are the guided cruise missiles and ballistic missiles of the rocket force. These would be used to snipe hardened military targets, storehouses, airbases, as well as heavier dual use infrastructures such as dams, bridges and so on.
The final tier will consist of the munitions carried by stealth platforms as well as hypersonic weapons of the rocket force. DF-17 has achieved a cost level similar to American Tomahawk, so fielding en masse isn't gonna be a problem. Hypersonic weapons can freely hit the highest value targets, such as leadership, major radar installations as well as any shipment of NATO "wunderwaffen" the moment it arrives for maximum demoralization. Due to the cost and maturity of the design, we will likely talk numbers of 500+ DF-17 at the lowest being available.
With this sort of approach to fighting, it only becomes a matter of matching targets with the number of munitions. KMT air defenses must activate and make themselves targets if they wish to fight the onslaught of the 1st and 2nd tier attacks, leaving them completely vulnerable to being sniped out by a hypersonic missile or stealth platforms.
If China goes at the KMT immediately and hard, the only rate that limits the matching of aim points with munitions is the speed which PLA can fire these munitions. The initial strikes will aim to leave nothing behind on Taiwan island, not for the separatists to use and not for the Americans to steal if Americans one day should make it on land. And to deliver these strikes will not necessarily take many weeks, more likely it might take days to a week.
I do not agree with waiting around in a blockade once the civil war has already gone hot. A blockade can be used in peacetime to enforce China's control, but if enacted during a shooting war without the PLA shooting back overwhelmingly, a blockade is a signal of weakness that causes unnecessary escalation.
Once firing begins, Beijing should continuously feint landings while the above strategy is used to flatten everything on the island. The goal is to force the KMT forces to retreat into inland strongholds with sheer volume of fire and cede control of the coasts to the PLA. By constantly scouting for places to land, sending unmanned vehicles, sending special forces and so on, China will be able to gauge the exact moment when the only remaining coherent KMT forces are all restricted to trenches and deep bunkers in the forests and mountains.
The sooner ground control has been established, the lower the risk for foreigners using the volatile situation as an excuse to invade. It also reduces the strain on the local civilians.