This study aims to determine the effect of the Montessori method in improving understanding of mathematical concepts for early childhood especially ages 4-6 years in the city of Banjarmasin. This study uses a quasi-experimental method. Introduction to mathematics in the experimental class uses the Montessori method, while the control class uses conventional methods. Hypothesis test results show that the implementation of the Montessori method has significant effects in improving children's learning outcomes in data and graphic material in the experimental class. The average value of N-Gain in the experimental class was 95.61. The average value of the N-Gain control class is 34.56. The Montessori method is 61.05% more effective in improving children's learning outcomes in material data and graphics than the conventional method. Similar to the results of hypothesis testing on the ABCDE-ABCDE pattern material, the implementation of the Montessori method has a significant effect on improving the learning outcomes of children in the experimental class. The average N-Gain of the experimental class is 87.50. The average value of the control class N-Gain is 31.24. The Montessori Method 56.26% is more effective in improving children's learning outcomes on material patterns than conventional methods.