The low ability of mathematical generalization of the students of SMPN 17 Bandar Lampung is because most of the students still have difficulty in drawing a conclusion from the material they have acquired and the lesson learned process of generalization aspect or less precisely the instructional model used. The purpose of this research is to know the influence of learning model of Peer Led Guided Inquiry and learning, discovery learning model with the ability of mathematical generalization. The measured variable from this research is the generalization ability in mathematics learning with the subject of algebraic factorization. This research is a quasi experimental research (quasi experimental research), where the subject of the study is students Class VIII students SMPN 17 Bandar Lampung. Data collection technique used is a test of mathematical generalization ability. The data analysis technique used analysis of variance (ANAVA) of one road with cell unlike the significance level α = 0,05 and as requirement analysis, that is a normal distribution population with liliefors model and homogenous population with Bartlett Test. Based on the result of research using ANAVA one way with cell unequal with significance level α = 0,05 it is obtained that 〖F〗 _count> F_tabel IE 7,5105> 3,0973, hence the decision of Ho test is rejected. So it can be concluded that there is influence of learning model Peer Led Guided Inquiry, learning model of Discovery Learning, Conventional learning model to ability of mathematical generalization of learners. The test proceeds with a double-compression test using Scheffe ' methods to see which has a significant effect on mathematical generabilizability. The calculation results show that of the three treatments, the learning model of Discovery Learning which gives a significant influence on the ability of mathematical generalization.