A first version of an instrument to assess science conceptual knowledge in pre-service primary teachers is elaborated and validated. A multiple-choice test with 50 items was proposed. Four content areas of science from the curriculum of Primary Education in Spain were considered: natural environment and its conservation, biological diversity, health education and the human body, matter and energy. A sample of 83 pre-service teachers from a spanish university participated in the statistical validation process. Both external consistency (Test-Retest correlation) and internal consistency (Kurder-Richardson's 20th formula) were above 0.7. Item difficulty and discrimination indices were analysed as well as distractor election. Although some adjustments should be made, indicators suggest that the instrument is appropriate for its purpose. Furthermore, remarkable results about participants' poor conceptual mastery and misconceptions were identified. Implications of these findings are discussed.