This paper estimates hedonic price models for a sample of 610 vacation rental houses located in the vicinity of a lake and four-season ski-golf resort. Hedonic semi-logarithmic regression models are estimated for peak summer and peak winter rentals for 2008. The regression estimates for weekly rental prices are conditioned on explanatory variables for house size, house quality, and recreation features including lakefront proximity and ski-slope access. The estimates are used to obtain percentage effects and implicit monetary values, and provide evidence that access to recreation opportunities is reflected importantly in rental offers. Evaluated at the sample means, lakefront locations have a rental premium of about $1100-1200 per week and the premium for ski-slope locations is $500-600 per week. Although there is evidence of positive spatial correlation in the OLS residuals, estimation of spatial models does not result in substantial changes in most coefficient magnitudes or standard error estimates.