First, the *good* answer is just a cleaned up version of your solution -- the main difference is that I'm leveraging the ability of Python `for ... in` loops to iterate through a string. The *better?* answer uses another optimization (or is it really optimized?) -- try and understand what it's doing.

findoriginal.py

# return the first non repeated character in a string

def find_original(string):
	count_dict = {}
	order = []
	for char in string:
		if char in count_dict:
			count_dict[char] += 1
		else:
			count_dict[char] = 1
			order.append(char)
	for x in order:
		if count_dict[x] == 1:
			return x
	return None

# good
def find_original(s):
    counts = {}

    for c in s:
        counts[c] = counts.get(c, 1)

    for c in s:
        if counts[c] == 1:
            return c

    return None

# better?
def find_original(s):
    positions = {}
    counts = []

    for i, c in enumerate(s):
        if not c.isalpha():
            continue

        positions[c] = positions.get(c, len(counts))
        position = positions[c]

        if len(counts) <= position:
            counts.append((c, 1))

        else:
            count = counts[position]
            counts[position] = (count[0], count[1] + 1)

    for c, count in counts:
        if count == 1:
            return c

    return None

tests = [
    ('catattack', 'k'),
    ('banana', 'b'),
    ('racecar', 'e'),
    ('order rowena', 'd')
]

for s, answer in tests:
    value = find_original(s)
    assert value == answer, "{0}: expected {1}, got {2}".format(s, answer, value)
    print "{0!r} -> {1!r}".format(s, value)

You may want to look up dict.get() to see what it does: http://www.tutorialspoint.com/python/dictionary_get.htm