The whole point of HA is to power up VM's after an ESX host crashes, its an emergency procedure which is not a per VM thing really. It's more of a per ESX host thing. What are you trying to test? Is it possibly you might be thinking of DRS? That works on a per VM level and you can test it without rebooting a VM or a host and you can also set per VM options for DRS which will migrate a VM from one host to another under certain resource constraints.
You can see all these options by selecting the cluster in the hosts and clusters view then select edit settings to see all these options. Otherwise the suggestion of putting one VM on one host and rebooting that ESX host is the only way to test HA (ie the VM will power down along with the ESX host then HA will power that VM up on another remaining host, that is what HA should do if working correctly).
Hope that helps,
Hanna
---
VCP2, VCP VI3, VCP vSphere