The purposes of this experimental study with pretest-posttest control group design were to examine the effect of community-based recreation program (CBRP) in developing of self-directed learning skills (SDLS) among teacher candidates and to evaluate their satisfaction of CBRP. The samples were 60 teacher candidates from Srinakharinwirot University. They were divided equally into two groups (experimental and control group) by using multi-stage sampling. Research instruments were: a CBRP for SDLS development of teacher candidates, which face validity was approved by five experts; and two questionnaires designed by researcher to evaluate teacher candidates' satisfaction of CBRP and assess their SDLS. Both questionnaires using IOC = 0.8-1.0 on all item with five experts also approved the content validity. Cronbach's Alpha on all item were 0.80 or higher respectively for coefficient of reliability. Data was analysed by means, S.D., and dependent t-test, independent t-test, and effect size (Cohen's d). The results indicated that: 1) after the CBRP was completed, the experimental group scored better in both overall and each item of SDLS than before joining the program with significant differences of 0.05, 2) there was no significant difference in the overall and each item of SDLS in the control group between before and after the CBRP was completed; 3) after the CBRP was completed, the experimental group SDLS's scores were better in both overall and each item than control group at the significant differences of 0.05. Finally, the effect size (Cohen's d) of this study indicated that the experimental group is large enough to have an influence on the population (d = 18.52); 4) the experimental group scored the highest on the satisfaction evaluation for the CBRP.