We present a smartphone-based bimodal device (SBBD) combining fluorescence imaging and spectroscopy for a noninvasive, portable, cost-effective, and real-time diagnosis of oral precancer. The device employs a 405 nm laser excitation to capture native fluorescence from intrinsic biomarkers, including flavin adenine dinucleotide (FAD) and porphyrin, across 136 subjects representing normal, OPMD, and OSCC. The fluorescence imaging module identifies regions of interest (ROI) based on the red-to-green band ratio (porphyrinFAD), while the spectroscopy module confirms findings through multiple point measurements. A 2D convolutional neural network (CNN) classifies normal tissues, oral potentially malignant disorders (OPMD), and oral squamous cell carcinoma (OSCC) with 97.04% accuracy, 96.13% sensitivity, and 97.73% specificity. Fluorescence spectroscopy, enhanced by an artificial neural network (ANN), achieves 97.44% accuracy, 95.24% sensitivity, and 97.44% specificity. This bimodal approach effectively addresses the diagnostic gap that occurs when either spectroscopy or imaging is used independently for oral cancer detection and biopsy guidance.