Penelitian ini bertujuan menganalisis pengaruh Environmental, Social, and Governance (ESG), Return on Assets (ROA), Firm Size, dan Earnings per Share (EPS) terhadap nilai perusahaan yang diproksikan dengan Tobin's Q pada perusahaan indeks LQ45 periode 2021–2024. Penelitian menggunakan pendekatan kuantitatif asosiatif kausal dengan 27 perusahaan sampel yang dipilih melalui purposive sampling, sehingga diperoleh 108 observasi. Data dianalisis menggunakan regresi data panel dengan model Random Effect Model melalui perangkat lunak EViews 13 dan estimasi robust standard error. Hasil penelitian menunjukkan bahwa secara simultan ESG, ROA, Firm Size, dan EPS berpengaruh signifikan terhadap nilai perusahaan. Secara parsial, ESG dan ROA tidak berpengaruh signifikan, Firm Size berpengaruh negatif dan signifikan, sedangkan EPS berpengaruh positif dan signifikan terhadap nilai perusahaan. Nilai koefisien determinasi sebesar 35,21% menunjukkan bahwa variabel dalam model menjelaskan variasi nilai perusahaan, sementara sisanya dijelaskan oleh variabel lain di luar model. ABSTRACT This study aims to analyze the effect of Environmental, Social, and Governance (ESG), Return on Assets (ROA), Firm Size, and Earning per Share (EPS) on firm value proxied by Tobin's Q in LQ45 index companies for the 2021–2024 period. This research uses a causal associative quantitative approach with 27 sample companies selected through purposive sampling, resulting in 108 observations. The data were analyzed using panel data regression with the Random Effect Model through EViews 13 software and robust standard error estimation. The results show that simultaneously ESG, ROA, Firm Size, and EPS have a significant effect on firm value. Partially, ESG and ROA have no significant effect, Firm Size has a significant negative effect, while EPS has a significant positive effect on firm value. The coefficient of determination of 35.21% indicates that the variables in the model explain the variation in firm value, while the remainder is explained by other variables outside the model.