RIKA ROSTIKA. NIM 123223013. Mathematics Learning With Models Learning Analysis Design Development Implementation Evaluation (ADPIE) Into Improving Ability Students Understanding Math Concepts. Paper, Kuningan: Mathematics Education Program, Muhammadiyah STKIP Kuningan, 2016. This study aims to determine whether there are positive implications between or not students understanding of mathematical concepts with models learning Analysis Design Development Implementation Evaluation (ADPIE. The hypothesis of this study is, “the ability of students understanding of mathematical concepts than students using learning models Analysis Design Development Implementation Evaluation better understanding of mathematical concept than student using conventional learning”. This study was conducted at SMP Negeri 1 Lebakwangi in the second semester of academic year 2015/2016. The selection of a simple random sample is taken to determine the experimental class and the control class. The sample used in eksperimental class as many as 30 students and as many as 29 control class students. Instrument to determine the understanding of mathematical concept by using a test essay, while to determine students responses to the Analysis Design Development Implementation Evaluation learning models by using a questionnair. The instrument tested prior to the 22 student in the population other thab the study sample was obtain 7 valid item description of 10 question. From the results of the test comprehension of mathematical concepts derived average value of the initial test experiment class 18,43 and control class 17,57. While at the end of the test the average value of the experiment class 44,13 and control class 32,33. Analysis using t test for statistical hypothesis testing. From these calculations obtained tcount = 0,589 for the test of equality of two mean, then consulted on ttabel at 5% significance level and 57 degrees of freedom, the value ttable = 2,003 because ttable < tcount < ttable = (2,003 < 0,589 < 2,003) then H0 is accepted, it means there is no difference between the mean value of the initial test experimental class and control class. While the difference in the two mean test values obtain tcount = 3,827 then consulted on ttable at 5% significance level and 57 degress of freedom, the value ttable = 2,003 because tcount > ttable (3,827 > 2,003) then H1 is accepted, it means the ability of students understanding of mathematical concepts that using models Analysis Design Development Implementation Evaluation better the understanding of the ability of students to use mathematical concept of conventional learning.