Obstructive sleep apnea (OSA), where the upper airway collapses repeatedly during sleep due to inadequate dilator muscle tone, is challenging to treat as current therapies are poorly tolerated or have variable and unpredictable efficacy. We propose a novel, optogenetics-based therapy, that stimulates upper airway dilator muscle contractions in response to light. To determine the feasibility of a novel optogenetics-based OSA therapy, we developed a rodent model of human sleep-related upper airway muscle atonia. Using this model, we evaluated intralingual delivery of candidate optogenetic constructs, notably a muscle-targeted approach that will likely have a favorable safety profile. rAAV serotype 9 viral vectors expressing a channelrhodopsin-2 variant, driven by a muscle-specific or nonspecific promoter were injected into rat tongues to compare strength and specificity of opsin expression. Light-evoked electromyographic responses were recorded in an acute, rodent model of OSA. Airway dilation was captured with ultrasound. The muscle-specific promoter produced sufficient opsin expression for light stimulation to restore and/or enhance electromyographic signals (linear mixed model, F = 140.0, p < 0.001) and induce visible tongue contraction and airway dilation. The muscle-specific promoter induced stronger (RM-ANOVA, F(1,8) = 10.0, p = 0.013) and more specific opsin expression than the nonspecific promoter in an otherwise equivalent construct. Viral DNA and RNA were robust in the tongue, but low or absent in all other tissues. Significant functional responses to direct optogenetic muscle activation were achieved following muscle-specific promoter-driven rAAV-mediated transduction, providing proof-of-concept for an optogenetic therapy for patients with inadequate dilator muscle activity during sleep.