Osteoporosis is a common disease characterized by compromised bone strength predisposing a person to an increased risk of osteoporotic fracture (OF). Recently, extensive efforts have been made to identify candidate genes underlying osteoporosis by the use of surrogate phenotypes, such as bone mineral density (BMD) and bone geometry. Among them, BMD is a suitable choice if we aim to classify the role of biological pathways for bone strength and to understand the bone conditions in the development of osteoporosis. However, evidences show that the genetic correlation between BMD and OF is very limited. In this review, we are mainly concerned with an important issue, i.e., phenotype choice in osteoporosis genetic research. For clarity, we address this issue with several arguments, and comments are made on most representative literature.