Gene expression programming (GEP) is capable of solving many prediction, classification, and optimization problem effectively. It uses a fixed-length chromosome representing a set of equations. However, the chromosome length significantly affects the algorithm’s performance. Different problems may require varied chromosome lengths to achieve good results. Only a few studies have been conducted to deal with chromosome length in GEP. Therefore, this study aimed to develop an adaptive GEP to find proper chromosome length during the evolutionary process. The study proposed that the chromosome length may be varied for each individual instead of using the same length in the population. The evolutionary process would adjust the chromosome length and the chromosome with proper length will tend to survive. Furthermore, the study proposed a contraction operator that could delete or insert an allele in the chromosome to make it short or extended. This operator is expected to adjust the chromosome length to its optimal. A special slice crossover was also proposed to accommodate the crossover between parents with different chromosome lengths. The proposed algorithms’ performance was investigated by solving three symbolic regression problems. Additionally, the performance was compared to related previous gene expression programming algorithms.