As an integral part of the decentralized finance (DeFi) ecosystem, decentralized lending platforms (DLPs) have gained massive traction with the recently revived interest in blockchain technology. However, with the traction and the novel services that are being emerged in the DeFi space, several interesting security vulnerabilities and attacks have been observed in the last few years. Oracle manipulation attacks have been witnessed numerous times on DLPs, and in this paper, we aim to secure DLPs from these attacks. We propose an algorithmic solution called SecDeLP , that secures a general DLP against oracle manipulation attacks that are performed using flash loans. We provide a theorem to prove that if certain conditions are satisfied on some specific system and input parameters then oracle manipulation attacks using flash loans must be reverted, hence preventing the attack. Furthermore, we present a practical analysis using empirical data and show that the constraints used in our solution are reasonable. Specifically, we introduce a new input parameter in the SecDeLP algorithm that is required for each crypto-asset available on the DLP. Next, using the past three years of market price data of several crypto-assets, we illustrate safe-touse values for this input parameter. We show that our requirements on this parameter are satisfied with a high degree of confidence. We also show that the cost overhead for implementing SecDeLP is minimal and practical.