In Veracruz, Mexico, it is challenging to visually identify endemic tropical timber species such as mahogany (Swietenia macrophylla), cedar (Cedrela odorata), parota (Enterolobium cyclocarpum) and rosewood (Tabebuia rosea) due to their highly similar textures. Reliance on visual experience alone can compromise the authenticity, quality and traceability of these species, which are critical aspects for the operation and sustainable management of small fine furniture industries and cabinetmaking workshops. This study presents an automated method based on the optimized combination of multiple local binary patterns (MLBP) and grey-level co-occurrence matrices (GLCM) for efficiently identifying tropical timber species. Five key texture metrics—entropy, energy, contrast, variance and homogeneity—were extracted and their discriminatory capacity was evaluated using rigorous statistical analyses. This highlighted significant results in the analysis of variance (ANOVA), with p-values below 0.05 for all metrics. Entropy showed the strongest results, with an F-statistic of 41.33 (p = 0.0018). The results indicate a statistically robust separation between species, as evidenced by comparative boxplots showing clear differences in metric distributions. This approach does not rely on supervised classifiers or large labelled databases, facilitating its application in real time and in practical environments. The findings demonstrate the method’s potential for developing integrated systems for automatically identifying tropical woods in real time that are optimized to operate on hardware with limited resources.