This research purpose to describe the effect of cooperative learning model type TPS and the beginning ability toward the prose research ability university student program study education of language and Indonesian literature STKIP YPM Bangko. This research using quasy experiment with factorial 2x2 design. The population is all of five semester student with class A as control class class B as experiment class. Data analysis technique used Anava one way and Anava two way, the calculation helping with SPSS 16 program and Microsoft Office Excel program. Based on the data analysis, the results are, ( 1) there is a different of the prose research ability between using cooperative learning model type TPS and conventional model (Farithmetic > Ftable = 22,443 > 4,00), (2) there is a different of the prose research ability between using cooperative learning model type TPS and conventional model for student with high beginning ability (Farithmetic > Ftable = 12,964 > 4,18), (3) there is a different of the prose research ability between using cooperative learning model type TPS and conventional model for student with low beginning ability (Farithmetic > Ftable = 33,577 > 4,21), (4) there is not a interaction between learning model and beginning ability toward the prose research ability. Based on the results, it can be concluded that the prose research ability student using cooperative learning model type TPS is higher than the student using conventional model, there is not a interaction between learning model and beginning ability toward the prose research ability.