I do not know of a way to enforce constraint naming conventions using SQL Server 2008 Policy-Based Management. And although you can use DDL triggers, you will have to write code to check for name use as not all constraints require a name when adding to the columns directly in the CREATE TABLE or ALTER TABLE statements.

For example, you can add primary keys and defaults without a name as follows:

CREATE TABLE dbo.Table_1
   (
   id INT NOT NULL PRIMARY KEY,
   id2 INT NOT NULL,
   id3 INT NOT NULL DEFAULT (1)
   )  ON [PRIMARY]
GO  

The next example is a DDL trigger that will check for any constraints being added and print "CONSTRAINT!" whenever a constraint is added to a table (new or altered):

CREATE TRIGGER trTableCheckConstraint
ON DATABASE 
FOR DDL_TABLE_EVENTS
AS 
   DECLARE @x XML, @sql NVARCHAR(MAX)
   SET @x = EVENTDATA()

   SET @sql = UPPER(@x.value('(/EVENT_INSTANCE/TSQLCommand/CommandText/text())[1]', 'nvarchar(max)'))

   IF (@sql LIKE '%CONSTRAINT%' OR @sql LIKE '%DEFAULT%' OR @sql LIKE '%PRIMARY%')
   BEGIN
       PRINT 'CONSTRAINT!'
   END 
   
GO 

In order to enforce a name, however, you will have to do some string parsing (RegEx would be ideal for this of course, if you are so inclined to add a CLR-based Scalar User-Defined Function), as the CommandText will contain the actual SQL statement that was executed. You will need to check in the CommandText for all types of constraints, preferably in a UDF, and rollback if no name is used, or if the name does not comply with your naming standards. And the initial check may return false positives (such as a column with the word 'Primary' in its name), so be sure to do due diligence when validating the constraints.