Abstract: To cope with the problem of controlling population diversity in gene expression programming (GEP), an adaptive population diversity tuning algorithm is proposed. A weighted measurement for population diversity is designed. The impact in terms of selection, crossover, and mutation operators on population diversity is analyzed in detail. A diversity algorithm for initial population (DAIP) maximizing the initial population diversity is proposed as well. Aiming to appropriately maintain the population diversity and achieve high evolution efficiency, adaptive crossover and mutation operations are developed and an adaptive population diversity tuning algorithm (APDTA) is developed. Experiments show that APDTA is efficient and effective.