Melon farming in Malaysia requires careful management of soil moisture and temperature to ensure healthy growth and high yields. Previous studies have shown that surface temperature influences these soil parameters and that real-time monitoring enables smarter irrigation decisions. This study investigates how ambient temperature affects soil moisture and soil temperature, in order to support data-driven irrigation strategies in tropical greenhouse agriculture. The study was conducted over four months in a controlled greenhouse in Terengganu, Malaysia, using six sensor stations. Surface temperature, soil temperature, and soil moisture were recorded every 30 min. After data cleaning and normalisation, quantile regression was used to assess the effect of surface temperature on soil parameters across different moisture levels. The quantile regression model at the median (τ = 0.5) revealed a statistically significant positive relationship between surface temperature and soil moisture (coefficient = 0.25926, p < 0.001, t = 21.51), with a low standard error (0.01205), indicating high model precision. In contrast, the relationship between surface temperature and soil temperature was weak and not statistically significant (coefficient = –0.08152, p = 0.09875), suggesting that soil temperature is buffered from ambient fluctuations. These findings confirm that quantile regression offers a more accurate understanding of soil moisture dynamics compared to traditional linear models. Integrating additional variables such as air humidity and atmospheric pressure could improve predictive irrigation strategies.