When I tried to skip "ACCEPT-ENCODING" of header, I found the behavior was not right.
I think the issue is the following two "if" in _send_request of httplib.

    def _send_request(self, method, url, body, headers):
        # Honor explicitly requested Host: and Accept-Encoding: headers.
        header_names = dict.fromkeys([k.lower() for k in headers])
        skips = {}
        if 'host' in header_names:
            skips['skip_host'] = 1
        if 'accept-encoding' in header_names:
            skips['skip_accept_encoding'] = 1

Please explain what the wrong behaviour that you see is, and what you expect the right behaviour should be.

That code is intended to either keep any user-supplied Accept-Encoding header field, or send “Accept-Encoding: identity” if the field is not supplied.

If you are looking for a way to avoid adding this field entirely, see the lower level putrequest() and related methods. This is documented behaviour: <https://docs.python.org/2.7/library/httplib.html#httplib.HTTPConnection.putrequest>.

Sorry, I thought I misunderstood the meaning.

I want no "ACCEPT-ENCODING" even "ACCEPT-ENCODING: identity".
I tried to modify the code from 
    if 'accept-encoding' in header_names:
to
    if not 'accept-encoding' in header_names:

The http request will be no "ACCEPT-ENCODING".
This is what I want.

As Martin has stated, this can be achieved with putrequest as explained in the doc:

https://docs.python.org/3.10/library/http.client.html#http.client.HTTPConnection.putrequest