If I have a LAN with 1000 PC from which class I should use IP addresses ?

you should use Class B Private IP Address. Because in Class C, you will get maximum254 and in Class A private ip address. In class B you will get maximum65534 ip address. for1000 pc your ip address range would be172.16.0.1-172.16.3.254.

you can use any classes (excluding D and E).

as we have the tools known as " Supernetting  and subnetting "

;)

class A ip address scheme is the best approach as it will provide more flexibility and segregation for future scalibility.

In  Class B address you will achieve  all you need.256*256=65,536 excluding the broadcast addresses.

Choose any class: A,B,C with the subnet mask /22

you can use A or B both. depends how you do subnetting.

Class B.. but with subnetting you can also Class B or even A..

LAN with1000 PC will use Class B.

classB ,THE Network with CIDR /22 ,use subneting 

Local Area Network it cover the rane of2km. if we use router and switch then subneting mask  get automatic ip address for all pc. 

You can use Class B(4 IP addresses)

But if i were you i would opt classless IP ranges( Supernetting & Subnetting)

- There is much less waste of IP addresses. as you can define your own range of IP addresses.

- It will be easier to configure routing if you are in a large enterprise.

- If you firm has branch offices spread out, it will help you create your own classless network ranges.

- Good in security.

- VPN, Vlans, routing etc will minimal and less heck tick.

- Only disadvantage over classfull IP address is that it is slightly more CPU intensive than the latter (But the CPU consumption i am referring is very minimal, unrealistic and negligible).

- These are more difficult to understand and configure properly. Once you get over it , you will increase your state confidence and Knowledge in Networking as an whole.