This study investigated how students‟ motivation in Chemistry is affected through the use of Programmed Instruction in abstract topics. The research location was Butere sub-county, Kakamega County, Kenya. Quasiexperimental research design was used to implement the study, using Solomon four-group as a model. A group of 300 form two students, who were purposively sampled from a target population of 841 students, was used. Those who were assigned into experimental groups received their instruction using Programmed Learning Software, while their counterparts in control groups were taught conventionally. Focus was on the topic “structure of the atom and the periodic table” because it is abstract in nature and many students perform dismally in it during national examinations. The Students‟ Motivational Level Determination Questionnaire (SMLDQ) was designed by the researcher and assessed for its validity and reliability, then used to collect raw data, which was analyzed both descriptively (using mean and standard deviation) and inferentially using one-way ANOVA at α=0.05. Results revealed that Programmed Instruction was superior to the conventional approaches because the sampled students‟ pre-test motivation scores were statistically similar (ME2=56.8, SD=14.2, MC1=53.7, SD=14.8, t(299)=-4.13, p=.272] but significantly different in the post-test, in favour of the experimental groups [ME1=66.8, SD=11.9, ME2=66.8, SD=13.8, MC1=56.0, SD=16.3, MC2=56.3, SD=14.9, F(3,296)=15.8, p<.001]. These findings have instructional implications in science education.