Check out DBMS_PARALLEL_EXECUTE.

But the first thing I would check out is - often Java uses multithreading because of processing in a row by row basis. If you can change that to set-based operations via SQL, it is common that you do not need the complexity of multithreading at all.

Here's a trivial demo of that....we start with a routine to "fix" a row value, we need to call it 1000000 times so we're quick to think of multi-threading solutions, when just an update would have been fine.

SQL> create table t as select rownum pk, d.* from dba_objects d,
  2   ( select 1 from dual connect by level <= 20 );

Table created.

SQL>
SQL> create index ix on t ( pk );

Index created.

SQL>
SQL>
SQL> create or replace
  2  procedure fix_row(p_idx int, p_new_val varchar2) is
  3  begin
  4    update t
  5    set    owner = p_new_val
  6    where  pk = p_idx;
  7  end;
  8  /

Procedure created.

SQL>
SQL> set timing on
SQL> begin
  2  for i in 1 .. 1000000
  3  loop
  4    fix_row(i,upper(i));
  5  end loop;
  6  end;
  7  /

PL/SQL procedure successfully completed.

Elapsed: 00:00:36.10
SQL>
SQL> update t
  2  set    owner = upper(pk)
  3  where  pk <= 1000000;

1000000 rows updated.

Elapsed: 00:00:02.52
SQL>
SQL>