Alter table syntax

Assuming you have NO FK's referencing the table...

CREATE TABLE NEWTABLE (COLNAME INT IDENTITY(1,1), NEXT_COLUMN .....)

SET IDENTITY_INSERT NEWTABLE ON
INSERT INTO NEWTABLE (LIST ALL COLUMNS) -- listing all columns but the identity
SELECT *
FROM TBL_NAME
ORDER BY {YOUR_IDENTITY}
SET IDENTITY_INSERT NEWTABLE OFF
IF ((SELECT COUNT(*) FROM NEWTABLE) = (SELECT COUNT(*) FROM TBL_NAME))
DROP TABLE TBL_NAME
exec SP_RENAME 'NEWTABLE', 'TBL_NAME

Without knowing your column layout it won't be exact, but above lists the general idea of the process you would follow to complete the task.  Please feel free to ask any questions.

Actually, in BrandonGalderisi's answer, if you want to preserve the values of COLNAME from the original table you would have to list the identity column in your insert statement.  I assume that is the intent since you set identity_insert on, else you would not need to do that and the values would just be reassigned starting at 1.
But overall he is right, if you made the change through Enterprise Manager and looked at the script it creates, it would do the same thing, create a new table with the modified column, copy the data into it, drop the original table and rename the new table back to the original name.

My interpretation that he was trying to change an existing field to an identity was that he wished to retain the values.

BrandonGalderisi  is right.!!
i want to retain the id values...

But in his listing above, you are not guaranteed to retain the original values unless you can guarantee in the original table, they start at 1, increment by 1 and have NO Gaps.  In such a case then the line 'SET IDENTITY_INSERT NEWTABLE ON' is not needed, because you are not overriding the identity values the new table will be trying to assign.  To retain the id values from the original table if they are not this perfect, you have to specify the column in your insert and select clauses when copying the values to the new table.  Using BrandonGalderisi's example, change to the folowing
  Insert into NEWTABLE(COLNAME, NEXT_COLUMN, ...) -- listing all columns including the identity
  Select COLNAME, NEXT_COLUMN, ...  -- again listing all columns
  From TBL_NAME
  ... continue as he listed
I tested this to make sure I was not crazy,  if you run the above listing as is you get the following error:
Server: Msg 545, Level 16, State 1, Line 10
Explicit value must be specified for identity column in table 'NEWTABLE' when IDENTITY_INSERT is set to ON.
and it does not finish properly.

This is the code I ran to test and illustrate my point:
CREATE TABLE TBL_NAME (COLNAME INT, NEXT_COLUMN varchar(10))
insert into tbl_name values (1, 'test1')
insert into tbl_name values (3, 'test3')

select * from tbl_name

CREATE TABLE NEWTABLE (COLNAME INT IDENTITY(1,1), NEXT_COLUMN varchar(10))

SET IDENTITY_INSERT NEWTABLE ON
INSERT INTO NEWTABLE (colname, next_column) -- listing all columns but the identity
SELECT colname, next_column
FROM TBL_NAME
ORDER BY colname
SET IDENTITY_INSERT NEWTABLE OFF
IF ((SELECT COUNT(*) FROM NEWTABLE) = (SELECT COUNT(*) FROM TBL_NAME))
DROP TABLE TBL_NAME
exec SP_RENAME 'NEWTABLE', 'TBL_NAME'

select * from tbl_name
go
sp_help 'tbl_name'
go

This gives exact same values in resulting table, even with gap in identity and sets up the identity column.