Alternative splicing (AS) is significantly related to tumor development as well as a patient’s clinical characteristics. This study was designed to systematically analyze the survival-associated AS signatures in Lung adenocarcinoma (LUAD). Among 30,735 AS events in 9,635 genes, we found that there were 1,429 AS in 1,125 genes which were conspicuously related to the overall survival of LUAD patients. Then, according to the seven types of AS events, we established AS signatures and constructed a new combined prognostic model. The Kaplan-Meier curve results showed that seven types of AS signatures and the combined prognostic model could divide patients into distinct prognoses. The ROC curve shows that all eight AS signatures had powerful predictive properties with different AUCs ranging from 0.708 to 0.849. Additionally, the elevated risk scores were positively related to higher TNM stage and metastasis. Interestingly, AS events and splicing factors (SFs) network shed light on a meaningful connection between prognostic AS genes and corresponding SFs. Moreover, we found that the combined prognostic model signature has a higher predictive ability than the mRNA signature. Furthermore, tumors at high risk might evade immune recognition by decreasing the expression of antigen presentation genes. Finally, we predicted the three most significant small molecule drugs to inhibit LUAD. Among them, NVP-AUY922 had the lowest IC50 value and might become a potential drug to prolong a patient’s survival. In conclusion, our study established a potential prognostic signature for LUAD patients, revealed a splicing network between AS and SFs and possible immune escape mechanism, and provided several small-molecule drugs to inhibit tumorigenesis.