The main disadvantages of glucagon-like peptide 1 (GLP-1) are its rapid degradation and excretion. These bottlenecks can be overcome by lipidation or other structural modification. The aim of this study was to design a series of long-acting GLP-1 analogues based on our previously discovered Xenopus GLP-1 analogs (1–3). The structure-activity relationship around lipidated 1–3 derivatives (1a–3l) with respect to in vitro potency as well as protraction was firstly explored. Compound 3g was selected for further modification. The Gly2 of 3g was replaced with Aib2, and a lactam constraint between Glu16 and Lys20 (i to i + 4) was introduced to further improve the in vivo activity and stability, affording compound 4. The receptor activation capability and in vitro stability of 4 were better than 3g and liraglutide. In addition, the hypoglycemic and insulinotropic activity of 4 was significantly better than liraglutide in db/db mice. Moreover, the enhanced in vitro stability of 4 translated into improved in vivo pharmacokinetic profiles and a prolonged antidiabetic duration. Administration of 4 twice daily for one week in diet-induced obese mice caused a significant decrease in food intake, body fat and body weight. The five-week treatment study on db/db mice of 4 further demonstrated the therapeutic effects of 4 on body weight, HbA1c and glucose tolerance. These preclinical studies demonstrate the therapeutic potential of 4 for type 2 diabetes and obesity. The present study also suggests that combined lipidation and conformational constraint strategy has potential to be used for improving the therapeutic properties of peptides.