In this paper probabilistic models are developed from a model of malaria parasite population dynamics proposed by Macdonald  based on Galton-Watson branching and generating Negative Binomial distribution with maximum likelihood estimation, to test a model of malaria parasite extinction. Values of the model parameters are calculated using a samples of patients treated at the Jayapura General Hospital. The response of malaria parasites to chloroquine plus amodiaquine was examined with in- vivo tests. Samples were taken from 40 people, with 17 sensitive samples, 9 samples of degree of RI, 8 samples of degree of RII, and 6 samples of degree RIII. The data sample used in the model must satisfy the Negative Binomial distribution. Model construction approach includes: the basic model, study models, and mathematical calculations. A goodness of fit test was used to test the feasibility of the model. The model was focused on predicting the probability of the extinction of mutant malaria parasites resistant to chloroquine plus amodiaquine in a local population. Data analysis revealed the point of the probability of the extinction of resistant malaria parasites using Galton-Watson theorem is s = 0.8814. The probability of the extinction of malaria parasites resistant to chloroquine plus amodiaquine is φNB =0,8902.