Kepler found his laws in 1609 (first and second law)
and 1619 (third law). This was quite a while before
Newton's law of universal gravitation was known in 1687.
Hence Kepler could not derive his laws from Newton's gravitational law.
Actually it went the other way round: Newton could derive his
gravitational law from Kepler's laws.
Kepler did not use any physics to find his laws.
He had large amounts of data available from planetary positions
measured by Tycho Brahe.
These data contradicted the simple assumption of the planets
moving in circular orbits with constant speed. The differences
were not big, but significant (in case of Mars up to 9 %).
But knowing circular orbits are wrong does not give any clue
about which orbits would be the correct ones,
i.e. predicting the known the observational data.
So Kepler actually tried out several possible curves
(excentric circles, centered ellipses, excentric ellipses,
and probably other oval curves).
And after many failed tries he found that ellipses (described by the 3 laws)
did the best fit (i.e. correctly predicted the observational data).