The current study aimed to identify the appropriate method of teaching creativity through comparing computer-based instruction with problem-solving instruction. A sample of 50 male students studying in the second grade of junior high school was chosen and divided into two equal groups. Both groups filled out the Torrance Test of Creative Thinking (TTCT) in pre- and post-test stages. They also completed the Raven Standard Progressive Matrices (RSPM) as a covariate variable. The results of covariance analysis revealed that the mean value for creativity score was higher for the students provided with the computer-based method of instruction than that of their counterparts provided with the problem-solving method of instruction while the effect of intelligence was controlled. The scores obtained for the fluency component of creativity and those for flexibility, originality, and elaboration components also increased more through problem-solving and computer-based method of instruction, respectively. To sum up, employing computer-based method of instruction resulted in a further increase in the students' creativity.