在Procedure中p_receiver为收件人地址，多收件人以“,”（英文逗号）分割，
p_sub为邮件标题，p_txt为邮件内容（支持HTML）。代码如下：

  type address_list is table of varchar2(100) index by binary_integer;
  my_address_list address_list;
  ---------------------------------------分割邮件地址----------------------------------------------
  procedure p_splite_str(p_str varchar2,p_splite_flag int default 1) is
    l_addr varchar2(254) := \'\';
    l_len  int;
    l_str  varchar2(4000);
    j      int := 0; --表示邮件地址或者附件的个数
  begin
    /*处理接收邮件地址列表，包括去空格、将;转换为,等*/
    l_str := trim(rtrim(replace(replace(p_str, \';\', \',\'), \' \', \'\'), \',\'));
    l_len := length(l_str);
    for i in 1 .. l_len loop
      if substr(l_str, i, 1) <> \',\' then
        l_addr := l_addr || substr(l_str, i, 1);
      else
        j := j + 1;
        if p_splite_flag = 1 then --表示处理邮件地址
          --前后需要加上\'<>\'，否则很多邮箱将不能发送邮件
          l_addr := \'<\' || l_addr || \'>\';
          --调用邮件发送过程
          my_address_list(j) := l_addr;
        end if;
        l_addr := \'\';
      end if;
      if i = l_len then
        j := j + 1;
        if p_splite_flag = 1 then
          --调用邮件发送过程
          l_addr := \'<\' || l_addr || \'>\';
          my_address_list(j) := l_addr;
        end if;
      end if;
    end loop;
  end;
  ------------------------------------------------写邮件头和邮件内容------------------------------------------
  procedure write_data(p_conn   in out nocopy utl_smtp.connection,
                       p_name   in varchar2,
                       p_value  in varchar2,
                       p_splite varchar2 default \':\',
                       p_crlf   varchar2 default l_crlf) is
  begin
    /* utl_raw.cast_to_raw 对解决中文乱码问题很重要*/
    utl_smtp.write_raw_data(p_conn, utl_raw.cast_to_raw(convert(p_name||p_splite||p_value ||p_crlf, \'ZHS16CGB231280\')));
  end;
  ----------------------------------------写mime邮件尾部-----------------------------------------------------
  procedure end_boundary(conn in out nocopy utl_smtp.connection, last in boolean default false) is
  begin
    utl_smtp.write_data(conn, utl_tcp.crlf);
    if (last) then
      utl_smtp.write_data(conn, last_boundary);
    end if;
  end;
  ---------------------------------------------真正发送邮件的过程--------------------------------------------
  procedure p_email(p_sendoraddress2 varchar2, --发送地址
                    p_receiveraddress2 varchar2) --接受地址
   is
    l_conn utl_smtp.connection; --定义连接
  begin
    /*初始化邮件服务器信息，连接邮件服务器*/
    l_conn := utl_smtp.open_connection(p_server, p_port);
    utl_smtp.helo(l_conn, p_server);
    /* smtp服务器登录校验 */
    if p_need_smtp = 1 then
      utl_smtp.command(l_conn, \'AUTH LOGIN\', \'\');
      utl_smtp.command(l_conn, utl_raw.cast_to_varchar2(utl_encode.base64_encode(utl_raw.cast_to_raw(p_user))));
      utl_smtp.command(l_conn, utl_raw.cast_to_varchar2(utl_encode.base64_encode(utl_raw.cast_to_raw(p_pass))));
    end if;

    /*设置发送地址和接收地址*/
    utl_smtp.mail(l_conn, p_sendoraddress2);
    utl_smtp.rcpt(l_conn, p_receiveraddress2);

    /*设置邮件头*/
    utl_smtp.open_data(l_conn);
    /*设置日期*/
    --write_data(l_conn, \'Date\', to_char(sysdate-1/3, \'dd Mon yy hh24:mi:ss\'));
    /*设置发送人*/
    write_data(l_conn, \'From\', p_sendor);
    /*设置接收人*/
    write_data(l_conn, \'To\', p_receiver);
    /*设置邮件主题*/
    select replace(\'=?GB2312?B?\'||utl_raw.cast_to_varchar2(utl_encode.base64_encode(rawtohex(p_sub)))||\'?=\',utl_tcp.crlf,\'\')
    into p_subject from dual;
    write_data(l_conn, \'Subject\', p_subject);

    write_data(l_conn, \'Content-Type\', multipart_mime_type);
    utl_smtp.write_data(l_conn, utl_tcp.crlf);
    utl_smtp.write_data(l_conn, first_boundary);
    write_data(l_conn, \'Content-Type\', \'text/html;charset=gb2312\');
    --单独空一行，否则，正文内容不显示
    utl_smtp.write_data(l_conn, utl_tcp.crlf);
    /* 设置邮件正文
      把分隔符还原成chr(10)。这主要是为了shell中调用该过程，如果有多行,则先把多行的内容合并成一行，并用 l_splite分隔
      然后用 l_crlf替换chr(10)。这一步是必须的，否则将不能发送邮件正文有多行的邮件
    */
    write_data(l_conn, \'\', replace(replace(p_txt, l_splite, chr(10)), chr(10), l_crlf), \'\', \'\');
    end_boundary(l_conn);

    /*关闭数据写入*/
    utl_smtp.close_data(l_conn);
    /*关闭连接*/
    utl_smtp.quit(l_conn);

  end;
  ---------------------------------------------主过程-----------------------------------------------------
begin
  l_sendoraddress := \'<\' || p_sendor || \'>\';
  p_splite_str(p_receiver);--处理邮件地址
  for k in 1 .. my_address_list.count loop
    p_email(l_sendoraddress, my_address_list(k));
  end loop;
  /*处理邮件地址，根据逗号分割邮件*/
end;