This study aimed to determine the effect of PBL model on students’ cognitive achievement on high and low students’ problem solving abilities. This study was quasi experiment research by using posttest only non-equivalent control group design with 22 students of X(3) grader science class as control class (conventional model) and 21 students of X(4) as experiment class (using PBL model) at one of senior high school in Surakarta. Students’ problem solving abilities illustrated as abilities in four aspects: identification the problem, identification the cause of problem, create the method of problem solution, examine the result of problem solving by using 40 items of multiple choice question refers to Problem Solving Skill Test [12]. Students’ problem solving abilities measured before the research is done and divided into high and low categories. Students’ cognitive achievement illustrated as students’ concept understanding and measured by using essay test was given after the PBL model used. Instrument had been validated by expert judgment and students. Data analyze by using t-test. The result showed: 1) Sig. (0,00<0,05) Ho rejected, there was difference of students’ achievement between experiment class and control class; 2) Sig (0,683 >0,05) H0 accepted, there wasn’t difference students’ achievement with high problem solving abilities between low problem solving abilities.