Experimental probability is not necessary always same. For example, you may flip a fair coin 10 times and you probably would get 6 heads and 4 tails.
Then the experimental probability of getting a head would be 6/10 = 60%.
Now you can do 10 trials again and you probably would get different result such as 5 heads and 5 tails. In this case, the experimental probability of getting a head would be 50%, which is different. Experimental probability depends on a result of actual trials, so it may be different.
Theoretical probability is always same.
Hope this helps.